P. 1
ThomHartleStrategiesAnalysisCollectionVol2

ThomHartleStrategiesAnalysisCollectionVol2

|Views: 715|Likes:
Published by api-3763249

More info:

Published by: api-3763249 on Oct 16, 2008
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

03/18/2014

pdf

text

original

Thom Hartle Strategy and Analysis collection, Vol.

2

Thom Hartle Strategy and Analysis collection, Vol. 2
Article 1: “Putting the squeeze on volatility” (Active Trader, November 2001). “The bottom and the Bands” (Active Trader, February 2003). “Relative volume analysis” (Active Trader, July 2003). “Short-term oscillator opportunities” (Active Trader, September 2003). “Retracement tendencies” (Active Trader, March 2004). “Percent b” (Active Trader, April 2004). “Following the money: Quarterly momentum and ETFs” (Active Trader, June 2004). “The market facilitation index” (Active Trader, October 2004). “The squat bar” (Active Trader, January 2005). “HOLDRS: Stock groups in a single trade” (Active Trader, April 2005). “Dissecting T-note futures: Tendencies and characteristics” (Active Trader, July 2005). “Index vs. Index: The QQQQ/SPY spread” (Active Trader, September 2005). “Comparing moving averages” (Currency Trader, November 2005). “The ETF money trail” (Active Trader, December 2005). “Follow the leaders” (Active Trader, February 2006). “The intraday MFI” (Active Trader, September 2006).

3 9 12 16 20 24 28 34 37 40 43 50 58 62 70 77

Article 2:

Article 3:

Article 4:

Article 5:

Article 6:

Article 7:

Article 8:

Article 9:

Article 10:

Article 11:

Article 12:

Article 13:

Article 14:

Article 15:

Article 16:

TRADING Strategies

Putting THE SQUEEZE ON volatility
eople sometimes complain about excessive market volatility, but volatility is a trading necessity for one simple reason: Once a position is entered, the market has to move to a point where a reasonable profit is available. Low volatility and sideways markets result in small profits and more losses. High volatility and trending markets mean good trades and more of them. It follows that identifying when a market is moving from a low-volatility environment to a high-volatility environment would improve the odds of profitable trading. The BandWidth Indicator (BWI) — a derivative of Bollinger Bands — can be used for this purpose.

P

Bollinger Band background
Bollinger Bands, developed by John Bollinger, are price envelopes plotted two standard deviations above and below a moving average of the closing prices. (For more information, see “John Bollinger: Focus on the markets,”Active Trader, January/February 2001.) The price envelope essentially captures historical volatility — identifying boundaries price is likely to fluctuate between based on past volatility levels. Wide envelopes are indicative of high volatility (upward or downward price trends), and narrow envelopes signify low volatility (sideways price action). Experienced traders know that a lowvolatility market condition often precedes up or down price trends (see the discussion of flag patterns in this month’s Technical Tool Insight). Therein lies the opportunity: If you are patient and wait for a low-volatility situation, you can take advantage of an increase in volatility. Bollinger’s recognition of how the bands contracted and expand-

BY THOM HARTLE

Volatility is essential for profitable short-term trading. The BandWidth Indicator can help you identify points at which a market is poised to switch from a low-volatility to a high-volatility phase.

3

www.activetradermag.com • November 2001 • ACTIVE TRADER

FIGURE 1 THE BWI ed to signal a change in volatility led to the development of the BWI. The BandWidth Indicator (BWI) helps identify extreme volatility situations. The low BWI reading here reflects a tight trading range.
IBM (IBM), daily 125.00 120.00 115.00 110.00 100.00 95.00 90.00 85.00 80.00 BWI .30

Identifying volatility shifts
The BWI is the width of the Bollinger Bands expressed as a percentage of the moving average. It is particularly helpful for finding stocks trading in narrow ranges that may be ready to break out. The formula for the BWI is:
(Upper Bollinger Band - Lower Bollinger Band) Moving average of price

The lower the volatility, the closer the upper and lower Bollinger Bands will be, and this difference will constitute a smaller percentage of the middle Bollinger Band (i.e., the moving average of price). Very low BWI readings are often followed by volatility expansions. Figure 1 (right, top) is a daily chart with Bollinger Bands using the default values of 20 bars for the lookback period and two standard deviations for the bands. The bottom chart is the BWI. When the BWI falls to historically low levels (such as a six-month low in Figure 1), market volatility is exceptionally low and the “Squeeze,” as Bollinger calls it, is on. During a Squeeze, the market enters a sideways consolidation and the moving average levels, following the center of the price bars. In such conditions the Bollinger Bands narrow dramatically and the BWI drops to low levels. During the week of Feb. 20, the BWI dropped below .10, producing the lowest reading on the chart. At this point, you could draw in the support and resistance levels for the trading range, as shown. Breakouts of such trading ranges can signal new trends. In this case, IBM dropped through the bottom of the trading range, breaking support just below 110 and falling 20 additional points. Notice the break of the support level in this example coincided with the BWI turning up. Sometimes the BWI will move up in advance of the break of

.20

.10
18 26 2 8 Jan. 2001 16 22 29 1 5 Feb. 12 20 26 1 5 March 12 19 26

Source: CQGNet

FIGURE 2 EARLY WARNING During the trading range in mid-November, the BWI began to fall, dropping below .10. On Nov. 27, the prices tested the upper Bollinger Band, reversed and closed at the low of the day. The next day the market gapped down and closed at the low with higher volume. The BWI turned up.
Microsoft (MSFT), daily A 74.57 69.71 64.86 60.00 55.00 50.00 BWI

.40 .30 .20 A .10
1,000,000 50,000 16 23 30 1 6 November 13 20 27 1 December

Volume

Source: CQGNet

4

www.activetradermag.com • November 2001 • ACTIVE TRADER

FIGURE 3 INTRADAY TRADING Here, the Bollinger Bands and the BWI are set to a lookback period of 10 bars. Twice, the BWI dropped to the .10 level as the 30-minute bars were forming triangle formations. IBM dropped more than a point in both of these situations.
IBM (IBM), 30-minutes 117.00 116.00 115.00 114.00 113.00 112.00 BWI .40 .30 .20 .10
28-8:30 12:30 29-8:30 12:30 2-8:30 July 12:30 3-8:30 5-8:30

support or resistance, forewarning the direction of the break. For example, when a trading range in Microsoft began in mid-November 2000 (see Figure 2) the BWI began to fall, dropping below .10. On Nov. 27, price tried to push to new highs but closed at the low of the day. The next day (point A), the stock gapped down and closed near the low on higher volume. At the same time, the BWI turned up. Microsoft had not broken support or resistance yet, but the rise in the BWI indicated volatility was rising, and the weak price action and volume increase as prices moved down suggested a break of support.

Source: CQGNet

FIGURE 4 THE HEAD FAKE A Squeeze is sometimes followed by a false breakout. In this case, the BWI narrowed to below .10, and the market broke through the resistance level and penetrated the upper Bollinger Band. However, the stock then reversed and trended back down to make a new low in the downtrend.
IBM (IBM), daily 130.00 125.00 120.00 117.72 115.00 109.51 105.00 101.30 100.00 BWI .25 .20 ..15 .10
7 24 1 May 8 15 22 1 June 12 19 26 3 10 July 17 24 1 7 August

The longer the lookback period, the fewer signals will be generated. Short lookback periods will identify low-volatility situations more often.
Intraday trading
Intraday traders can use the BWI Squeeze as a setup for short-term trades. Figure 3 is a 30-minute chart with 10-bar Bollinger Bands/BWI. There are two instances where the BWI dropped to the .10 level as the stock was tracing out triangle formations. IBM dropped more than a point after breaking out of both these patterns. In this example, the lookback period was shortened because intraday activity tends to consist of short price trends, followed by consolidation, and new trends. The longer the lookback period, the fewer signals will be generated. Short

Source: CQGNet

5

www.activetradermag.com • November 2001 • ACTIVE TRADER

FIGURE 5 LONGER-TERM ANALYSIS lookback periods will identify lowvolatility situations more often. This monthly logarithmic scale chart of the Dow Jones Industrial Average illustrates how the index has been in a trading range for a number of years. BWI readings below .20 reflect the low volatility.
Dow Jones Industrial Average (DJIA), monthly
12,000 11,391 10,688 9,985 9,000 8,000 7,000 6,000 5,000 4,000

Head fakes
Bollinger cautions that BWI signals are subject to “head fakes” — when the BWI drops to a historically low level and the market moves through support or resistance, but suddenly reverses direction. Figure 4 shows a head fake. The BWI fell below .10 and the stock broke through resistance and penetrated the upper Bollinger Band. However, it then reversed and trended back down to make a new low. There are two ways to trade the Squeeze without getting burned. First, trade in the direction of the breakout with a tight trailing stop. That increases your chances of locking in a profit. (One possible solution of this type is to use Welles Wilder’s Parabolic SAR, which is a technique that trails a stop increasingly closer to price as time goes by. The Parabolic technique is included in many software packages and will be discussed in the December issue of Active Trader.) The second technique is to wait for the initial breakout to occur and watch the market to see if the trend from the initial breakout holds. If it does, commit to the trade; if it is a head fake, go with the reversal of the false breakout and take advantage of the fact that many traders will be on the wrong side of the real trend. Other exit strategies can be developed by using a maximum favorable excursion (MFE) analysis of 30 to 50 of the setups. MFE analyzes the typical maximum profit of a trading strategy (see, “On-target trading,” Active Trader, July 2001). That way, you can determine typical action after the Squeeze sets up and establish a preset target to exit a portion of your position. Also, consider taking partial profits when price tags the upper or lower Bollinger Band.

BWI .40 .20 .132
1997 1998 1999 2000 2001

Source: CQGNet

FIGURE 6 DOW JONES INDUSTRIAL AVERAGE (1991-1996) The BWI dropped to low levels during 1993 and early 1995. The dip during 1993 was followed by a 500-point rally — a 14-percent increase. The dip by the BWI during 1995 also was followed by a substantial rally.
Dow Jones Industrial Average (DJIA), monthly 6,000 5,500 5,000 4,500 4,000 3,500 3,000

2,500 BWI

Thinking big picture
The BWI can also be used for long-term analysis. For example, Figure 5 is a logarithmic scale (i.e., price movement shown as percentages, rather than

.40 .20
1991 1992 1993 1994 1995 1996

Source: CQGNet

ACTIVE TRADER • November 2001 • www.activetradermag.com

6

The greatest
FIGURE 7 DOW JONES INDUSTRIAL AVERAGE (1982-1986) The BWI dropped to low levels during the 1983 to 1984 trading range. Another major bull move followed.
Dow Jones Industrial Average (DJIA), monthly 2,250 2,000 1,750 1,500 1,250

strength of the BWI is it makes it easier to find low-volatility situations that can presage a new trend.
absolute point values) monthly chart of the Dow Jones Industrial Average that illustrates how this index has been in a trading range for the better part of the last two years. Figure 6 shows the BWI dropped to low levels during 1993 and early 1995. The low reading by the BWI in 1993 was followed by a 500-point rally — a 14percent increase. The next dip during 1995 was followed by another substantial rally. Figure 7 shows the Dow and the BWI during the 1983-1984 trading range. After the BWI bottomed in late 1984, a major bull market followed. However, sustained major trends don’t always follow such BWI readings. Figure 8 shows the tumultuous period of the late 1970s and early 1980s. The BWI dropped to low levels in late 1979. During the first quarter of 1980, the Dow tested both major resistance and support before breaking out above 900, only to move back below the original resistance level one year later. Finally, Figure 9 shows a head fake in mid-1968. The BWI dropped to low levels and the Dow broke through the 950 resistance level. However, the rally failed and the bear market ultimately resumed.

1,000

750 BWI .60 .40 .20
1982 1983 1984 1985 1986 1987

Source: CQGNet

FIGURE 8 DOW JONES INDUSTRIAL AVERAGE (1977-1981) During this tumultuous period of the late 1970s and early 1980s, the BWI dropped to low levels in late 1979. During the first quarter of 1980, the Dow tested both major resistance and support before breaking out above 900, only to move back below the original resistance level one year later.
Dow Jones Industrial Average (DJIA), monthly 1,050 1,000 950 900 850 800 750 700 BWI .40 .30 .20
1977 1978 1979 1980 1981

Making the most of the BWI
The greatest strength of the BWI is it makes it easier to find low-volatility situations that can presage a new trend. In addition, the BWI can help you avoid marginal trades (trades with a high BWI reading). Wait for the volatility to settle down and a trading range to establish itself, and then trade according to your rules.

Source: CQGNet

7

www.activetradermag.com • November 2001 • ACTIVE TRADER

FIGURE 9 DOW JONES INDUSTRIAL AVERAGE (1966-1970) The Dow made a head fake when the BWI dropped to low levels in mid-1968 and the Dow broke out through the 950 resistance level. However, the rally failed and the bear market ultimately resumed.
Dow Jones Industrial Average (DJIA), monthly 1,000 950 900 850 800 750 700 650 BWI .40 .30 .20
1966 1967 1968 1969 1970

In general, high BWI readings indicate the better part of the current trend is already over. Track the BWI to find trades that put you in the early phase of the trend. Keep in mind, though, that in the current market environment, stocks are much lower in price than they have been the last few years. As a result, the nominal moves will not be what you are accustomed to. After the Squeeze, you may want to use percentages — instead of points — for exit and stop-loss targets. To accompany his new book, Bollinger on Bollinger Bands (McGraw-Hill, 2001), Bollinger has created a Web site, www.bollingeronbollingerbands.com, which provides lists of stocks that meet his criteria for this technique as well as others he details in his book.

Source: CQGNet

ACTIVE TRADER • November 2001 • www.activetradermag.com

8

TRADING Strategies

The BOTTOM and the BANDS
The W-bottom pattern reveals a great deal about market action — especially when it’s used in conjunction with Bollinger Bands.

FIGURE 1 WATCH THE VOLUME A volume spike accompanied a new low at point A. However, a retest of the low at point D occurred on light volume, a sign the market was ready to move up.

BY THOM HARTLE

Coach Inc. (COH), daily

32.50

30.00

T

he technique for trading the W-bottom pattern, as described by John Bollinger in his book Bollinger on Bollinger Bands, blends classic technical analysis and modern indicator-based analysis. It’s classic because the W-bottom concept consists of a double bottom supported by traditional volume characteristics: heavy volume on the first low, a price recovery and a retest of the low on lighter volume. The high volume on the first low indicates a climactic selling spree: Sellers are jumping ship, most likely because of a news event. However, this low marks a bearish sentiment peak. The retest — when the market fails to make new lows — occurs on low volume, which suggests there are no more shares to sell at this price level. As a result, the market is set up to rally if there is any good news on the horizon.

B

E

F
Resistance

27.50

25.00

A
Volume Volume spike

C

D

22.50

2M

1M

0 16 Source: eSignal 23 30 October 7 14 21

The approach also takes advantage of the more modern statistical price measurement of Bollinger Bands. Augmenting W-bottoms with Bollinger Bands enables you to determine whether price action is within a typical range. In addition,

Bollinger Bands can help find stocks that are setting up W-bottoms. The following trading approach combines the implications of how price behaves relative to 20-day Bollinger Bands with classic price-volume chart analysis.

9

www.activetradermag.com • February 2003 • ACTIVE TRADER

The classic
FIGURE 2 W-BOTTOM ON A WEEKLY CHART Regardless of the time frame, W-bottoms usually exhibit the same price-volume relationships. Long trades are signaled when price pushes above the resistance level of the W-bottom, which subsequently acts as support.
Microsoft (MSFT), weekly 65.00

approach is to wait until the market breaks above the resistance level established during the rebound following the first low. From that point on, a trailing stop is progressively raised to just under each successive support point in a market advance.
double-bottom patterns is to wait until the market breaks (or closes) above the resistance level established during the rebound after the first low, which in this case would mean buying on bar F. After that, a trailing stop is used to protect the position. The stop is moved up just under each support level as the market advances. Another good idea is to take partial profits if the upper Bollinger Band is tagged (which suggests price has stretched to the higher end of its expected range), and use a trailing stop for the remainder of the position. Figure 2 shows an example on a weekly chart. Bar A closed below the lower Bollinger Band on a substantial volume increase. The stock then recovered, clos-

60.00 Resistance

B D

55.00

50.00

C A
Volume Volume spike

Support

45.00

E

236M

0 March April May June July August Sept. Oct. Nov.

Source: eSignal

Price and volume: The W-bottom
In Figure 1, bar A closed below the lower Bollinger Band on a significant volume increase. This is evidence of extreme downside price movement. In this case, the stock began to recover the same day, rallying and closing at the high of bar A. The recovery process lasted three more days (bar B), with volume running less than 50 percent of what it was on the climactic low day. Then, after this short recovery period, the stock made another significant penetration of the lower Bollinger Band on increased volume (bar C). The stock subsequently traded within the confines of this day’s range (and on lower volume) for two days. These two days, though, were above the lower Bollinger Band. Bar D opened much lower, near the support level of the bar A low. However, instead of sellers entering the market, buyers stepped in and the stock closed near the day’s high on more volume than

the previous two days. This is part of the strategy template: The first low and close occur below the lower Bollinger Band, and the second test of the low is above the lower Bollinger Band. The following day, bar E gapped up on another volume increase (this occurred right at the resistance level established during the recovery of the bar after A). At bar F, the stock began the day testing the previous day’s low, but then rallied through resistance on low volume — a sign there were very few sellers. The close above resistance signaled higher prices, and the market gapped up again the next day. What about trade management? Bollinger recommends entering on the close of the first bar that 1) follows the test of the low and 2) closes near its high on increasing volume. In Figure 1, this would mean buying on bar D’s close and placing a stop-loss below the low. The classic approach for trading W- or

ACTIVE TRADER • February 2003 • www.activetradermag.com

10

A close outside a Bollinger Band does not indicate the market will simply retrace back to the moving average.
The stock did advance the following week (D), but it encountered sellers at the moving average and closed below the midpoint of the week on higher volume. This price-volume action suggested a test was still underway at this point. The stock traded lower the following three weeks (bars), each week closing toward the lower end of the weekly range on increasing volume. At bar E, the stock initially tested the previous week’s low, but it rallied to close at the top of the bar with the highest volume since the week of bar A. On heavy volume, the stock then punched through the resistance formed by the high volume. It then rallied two days and established a resistance level (point B) just below the moving average. The stock subsequently retreated, closing at or near the daily low of the next five bars (with one exception) on heavy volume, although all the closes were still above the lower Bollinger Band. Bar C opened very near the previous day’s closing price and traded toward the lower Bollinger Band, but encountering strong demand it eventually closed near its high. The volume was slightly less than the previous two days, but still heavier than average. The next day, the stock gapped up over the resistance formed by the high of bar B and closed at the high of the day. This bar was accompanied by volume greater than that of the mid-September spike-volume day, and it left a lot of sellers behind. The market continued to advance on increasing volume. Figures 2 and 3 use the same pattern and indicator analysis to examine the same data on different time frames. While the daily chart does not have the precise volume characteristics of the classic double bottom, the price and volume activity comprise part of the bottoming process shown on the weekly chart.

FIGURE 3 SAME STOCK, DIFFERENT TIME FRAME Two back-to-back weekly bars from Figure 2 turn out to form a W-bottom on the daily time frame. Although the volume characteristics are slightly different than those of a classic W-bottom, this formation functioned as a bottoming pattern within a bottoming pattern.
Microsoft (MSFT), daily 57.2676 55.0000 52.2930 50.0000

B

47.3185 Resistance 45.0000

Bollinger Bands: Trading tool and scanner
Bollinger Bands are a statistically derived measurement of extreme price action relative to average price action. A close outside the Bollinger Bands represents excessive price movement, but it does not automatically mean the market will retrace back to the moving average (or the opposite band). However, a valuable analytical template is created by adding the price and volume characteristics of a classic double bottom. Bollinger Bands enable you to screen for setups consisting of the combination of a close below the lower Bollinger Band occurring on volume greater than the 30-day average volume. This will produce a list of stocks with the potential to form W-bottoms. The support and resistance levels defined by the W-bottom pattern allow you to construct a set of procedures for managing trades once the pattern is complete.

A
Volume Volume spike

C
42.5000 Second highest volume 50M

19

26

3 9 September

16

23

30 7 October

14

21

28

0 4 November

Source: eSignal

ing (on declining volume) just above the moving average of the Bollinger Bands at point B. Next, the stock turned back again, and at bar C it closed at the upper end of that week’s range. At this point, a successful low-volume test appeared to be complete because the price held above the lower Bollinger Band. The market should be set to rally.
11

weekly reversal high at bar D; the resistance level subsequently functioned as a support level. A closer look at this chart indicates Bar E and the preceding bar may have formed a double bottom of their own. Let’s look at those two weeks using daily bars (see Figure 3). At bar A, the stock closed below the lower Bollinger Band on

www.activetradermag.com • February 2003 • ACTIVE TRADER

TRADING Strategies

Relative VOLUME analysis
Traders commonly use total volume to take the pulse of price moves, but when it comes to determining whether the bulls or bears are in charge of the market, up and down volume speak volumes.

BY THOM HARTLE

olume has always been a component of classic technical analysis. For example, one of the precepts of trend analysis is that volume should increase in the direction of the longer-term trend and decrease during countertrend moves. Expanding volume during a price move implies many market participants are driving the price action, confirming the trend. If volume does not increase, the move is suspect because of the lack of commitment implied by the weak volume. Although basic volume analysis can be useful, there’s more to volume than first meets the eye. End-of-day volume figures are comprised of the volume for stocks that are up (above the previous close), down (below the previous close) and unchanged on the day. More so than total volume, up volume and down volume can provide additional insight about the direction of the stock market.
12

V

Dividing the up-volume and downvolume numbers for each day by the total volume (not counting unchanged volume) gives the percentage of the day’s volume attributable to stocks that are trading up or down. These percentages aid analysis by revealing the degree

of buying or selling taking place at a specific time, such as when price is challenging a support or resistance level. The following examples analyze endof-day up-volume and down-volume statistics in the S&P 500 index (see Figure 1). Candlestick charts are used to

www.activetradermag.com • July 2003 • ACTIVE TRADER

highlight whether a particular day closed above the opening price (a white candle) or below the opening price (a black candle). Below the price candles are green and red volume bars that reflect the relative up or down volume.

Bar-by-bar breakdown
Figure 1 shows the percentage of volume that was up (green bars) or down (red bars) — whichever was greater relative to the combination of up and down volume for the day. FIGURE 1 RELATIVE VOLUME

The first basic principle of this analysis is that during an uptrend, the up-volume (green) bars will generally be higher than the immediately preceding down-volume (red) bars. In Figure 1, which is characterized by rising prices, the majority of the green volume bars are higher than the red volume bars, a clear sign of an uptrend. Bars 2 through 5 are all green, and bar 3 reached 90 percent, which means nine out of 10 shares traded that day were in stocks up on the day, a sign of strong

The green and red volume bars show the percentage of volume that was up or down, respectively, on a given day.
950 930 910 11 890 870 850 830 810 790 770 90% 80% 70% 60% 50% 40% 10/14 Source: Excel 10/21 10/28 11/4 11/11 11/18 4 1 2 3 5 Up /down volume 6 7 8 9 10 13 11 12 15 16 14 18 19 17 20 26 24 22 21 23 25 27 28 2930 33 32 31 2 1 3 4 5 6 7 10 8 9 12 13 14 15 S&P 500 index (SPX), daily 20 17 16 19 18 21 30 22 23 24 25 27 26 2829 32 31 33

demand. Candles 2-5 accounted for the largest uninterrupted up move on the chart. Candle 6 was an inside day that closed lower than the open. The down volume that day was just less than 85 percent of the day’s total up-down volume. That this down-volume bar was slightly lower than the previous up-volume bars suggests candle 6 is just a correction, or pause, in the uptrend. Had the red-volume bars been larger than the previous green-volume bars, a reversal would have been more likely. The market advanced further during candles 7, 8 and 9, but their up-volume bars were all lower than candle 6’s down-volume bar, and up-volume bar 9 was lower than previous up-volume bars 3 and 5. This means the up-volume percentage was dropping off and the higher prices were not attracting that much additional buying; the market was becoming more balanced, with many stocks up and many others down. The S&P made a new high (for the period shown on the chart) at candle 14, but this upside breakout failed and the index closed down for the day. However, its down-volume bar is lower than the previous up-volume bars. Candle 15’s low broke below the immediate support level established by bars 11, 12 and 13, but it closed back above this level; again, the down-volume bar was still lower than the most recent up-volume bar. The lack of high-percentage down volume in this context implied the longerterm uptrend is still intact. Generally, the down-volume bars would be starting to dominate in such a situation if the market was going to break down. However, volume bars 16 and 17 in this emerging trading range were dropping. Candle 18 was an up day, and its up-

ACTIVE TRADER • July 2003 • www.activetradermag.com

13

Relative volume reveals the degree of buying and selling taking place at a specific time, such as when price is challenging a support or resistance level.
volume bar was the highest in the previous eight days, a sign buyers were coming back into the market after the test of support. But as the market started to climb, the up-volume bars trended down: bars 19, 20 and 21 were all lower than bar 18. Conclusion: The higher prices were attracting sellers in the background. Candle 22 opened and reversed at the high of the recent trading range, closing below the lows of candles 19, 20 and 21. But more importantly, volume bar 22 was the highest of the last 12 volume bars, which meant sellers were asserting control in the market. The S&P subsequently tumbled to the support level (at candle 24) and the down-volume bars were higher than the recent up-volume bars. The high down-volume percentages at the test of support would indicate a strong likelihood the market would drop below this level. However, although candle 26 took out candle 24’s low, the market recovered. Also, the down-volume percentage for this bar was the lowest on the chart (barely above 50 percent), which means sellers were not participating aggressively at this level — they had already sold. This evidence points to a potential upside reversal. The next day, candle 27, marked another advance, with an up-volume bar substantially higher than the previous down-volume bar. Candle 28 also moved higher, but its volume bar was lower than that of bar 27. Candles 29 and 30 moved lower, but their down-volume bars were lower than the previous two
14

that up volume will dominate down volume in an uptrend. The key sign of an uptrend, though, is the relatively low down-volume percentage during corrections. Let’s now look at how the volume can reverse as the market forms a short-term top.

Reversal signal
Figure 2 (below) shows the S&P 500 during January 2003. Candles 1, 2 and 3 all had strong up-volume bars; the intervening down-volume bars were much lower, all of which underscored bullish price-volume action. However, as the market tested resistance around 930 after candle 3, the upvolume percentage dropped off, indicating selling was occurring in the background. Candle 4 closed below the low

up-volume bars, which meant the sellers were still not in control and this move was likely a correction in the overall uptrend. Candles 31 and 32 were bullish candles with relatively strong up volume (increasing up-volume bars). This chart supports the general idea

FIGURE 2 CHANGING OF THE VOLUME GUARD As the market tested resistance around 930, the up-volume percentage dropped off (between bars 3 and 4), hinting at background selling.
S&P 500 index (SPX), daily 940 930 920 910 900 890 880 870 860 850 840 1 90% 80% 70% 60% 50% 40% 1/6 Source: Excel 1/13 1/21 1/27 2 3 4 5 Up/down volume 1 2 3 5 4

www.activetradermag.com • July 2003 • ACTIVE TRADER

of the previous two days on a higher down-volume percentage than the three previous up-volume days — a sign sellers were coming into the market. FIGURE 3 BOTTOM REVERSAL

was trending lower, with the down-volume percentages dominating the up-volume percentages. On March 10 (candle 1), the down volume was above 90 per-

As the market bottomed out at candles 1, 2 and 3, the corresponding down-volume bars shrank.
S&P 500 index (SPX), daily 890 870 850 830 810 790 1 90% 80% 70% 60% 50% 40% 3/3 3/10 3/17 Up/down volume 4 2 3 5 1 2 3 4

5

day, though, the market started to advance, and by the end of the day the high up-volume percentage (over 90 percent) confirmed the up move and alerted traders to the potential for additional gains. However, notice the up-volume bars began to decline at this point, similar to the way the down-volume bars for candles 2 and 3 did when the market reversed to the upside. The two situations have an important difference, though: Candle 4’s up-volume bar was substantially higher than candle 2 and 3’s down-volume bars. You must be able to compare up-volume bars to down-volume bars (or vice versa) — not up-volume to up-volume or downvolume to down-volume — before drawing any conclusions about potential reversals. Therefore, until the market made a correction following the rally that began at candle 4, you must assume the trend is still up. The key to this analysis is that countertrend volume patterns indicate whether the trend is up or down. Low down-volume bars indicate the trend is up. Low up-volume bars indicate the trend is down.

Trade influence
Price-volume analysis is a useful barometer of the market environment. If your analysis in a market indicates the principles outlined here are in effect, consider executing your strategies in the direction of the trend indicated by the volume patterns — i.e., if the up-volume bars are green and the down-volume bars are red, long positions are likely to perform better than short trades. One thing to keep in mind is that it is always possible that news can trigger major buying or selling and reverse the trend. However, following the percentage of up-volume and down-volume relative to the total up-down volume during a trend (and at key support and resistance levels) can help clarify whether the bulls or the bears are dominating the market at a given moment.

Source: Excel

The market moved steadily lower over the next few candles on a rising percentage of down volume. Candle 5 was an up day, but the up-volume bar is lower than the previous down-volume bars — a negative sign. Subsequently, the market continued to trend down, with high percentage down-volume days and low percentage up-volume days. The final example illustrates the reversal of a downtrend. Figure 3 (above) shows the S&P 500 and its up and down volume percentages for the first few weeks of March 2003. From March 3 through March 12, the market

cent. However, candle 2 and candle 3’s down-volume percentages dropped even as price continued to fall. In other words, the lower prices were not bringing more sellers into the market. Candle 3 made a new low but reversed to close above its open, hinting the market was vulnerable to a short-covering rally. Candle 4 was a big up day. The upvolume bar was much higher than the down-volume bars of the two previous down days. Bar 5 is more ambiguous. The market closed above the 830 resistance level established during the previous week but the up-volume percentage was low, reflecting uncertainty. The next

ACTIVE TRADER • July 2003 • www.activetradermag.com

15

FUTURES

&

OPTIONS

Trading Strategies

Short-term oscillator opportunities
Exploring the relationship between an exponential moving average and an oscillator leads to new ways of defining trend changes and capturing price swings.
BY THOM HARTLE

A

lthough it sometimes seems as if trends consist of smooth, orderly up or down moves (especially if your trade is a loser), prices usually zigzag up and down even during strong trends. If the longer-term trend is up, price will rise from a shorter-term oversold state to an overbought state and back again, repeating this cycle until the market tops, then moves into a downtrend. The difficulty is that while these price swings are easily

identifiable on historical charts, there is rarely a regular time relationship between these swing peaks and troughs. Life would certainly be easy if a market would simply bottom, rise for five days and peak, fall for three days and bottom, and then rise for five days again, repeating this pattern indefinitely. The reality is the number of price bars between swing peaks and troughs throughout a trend is much more random. Nonetheless, some straightforward analysis makes it possible to identify trade opportunities using the concept that markets zigzag around their longerFIGURE 1 THE EMA-RSI CONNECTION term trends. To spot such opportunities relative to the Closes below the six-day EMA correspond to RSI readings below 50. trend, we must first select a method to define the trend. The moving average is probably the U.S. Treasury Bonds (US), daily most popular trend-defining tool, but it is not 112 the only one. Oscillators, which are calculaSix-day exponential moving average tions that highlight shorter-term price swings, are usually thought of as tools for identifying overbought/oversold conditions or diver111 gences, not as trend indicators. However, one of the most popular oscillators, the relative strength index (RSI, see “The relative strength index”), can be used with a 110 simple moving average to trade with the trend and to warn of trend reversals.
109 30-day simple moving average RSI A B C D E F G H I J 75 50 February 27 Source: CQGNet 3 10 18 24 25

Uptrends and downtrends
In technical analysis, as in most other areas of life, you always give up something to gain something else; for every advantage an indicator or approach possesses, it also has a drawback. For example, the longer a moving average’s lookback period (the number of price bars used to calculate the average), the less likely it will result in false signals (when price repeatedly moves above and below the average), but the more it will lag changes in market direction. Although there are several types

16

www.activetradermag.com • September 2003 • ACTIVE TRADER

FIGURE 2 THE TREND EFFECT The RSI peaks during this period were all below 75, while downward price thrusts pushed the indicator below 25.
S&P 500 E-Mini (ES), daily 1,140

of moving averages (simple, exponential, weighted, etc.), they all have one purpose: identifying the trend by smoothing price action and removing market “noise.” This allows traders to create simple definitions for trend direction, such as the one we will use here: If the moving average is rising, the trend is up; if the moving average is falling, the trend is down. Using the direction of the moving average to define the trend, we can then move on to the next step of qualifying the nature of intratrend price swings.

30-day SMA 1,120

1,100

108

106

Adding the oscillator
Most oscillators, in one way or another, measure the difference between prices over a short 75 A B C D E F RSI time period (e.g., three to 20 days). This 50 removes any longer-term trend influence and produces a horizontal picture of price zigzag25 ging above and below the trend, which is rep0 May resented by the oscillator’s median line (the 8 15 22 29 1 6 13 midpoint or “zero line”). Source: CQGNet “The trend-following RSI” (January 2003, Active Trader, p. 44) illustrated how crosses above and below the RSI median line are nearly equivalent to in the direction of the trend depending on the strength of that price crossovers of a moving average with twice the length of trend — readings will be skewed higher in uptrends and lower in downtrends. In this case, the low RSI values are around 40, the RSI lookback period. For example, Figure 1 (opposite page) shows T-bond futures which confirms the strength of the uptrend (which is also with a rising 30-day simple moving average (indicating the reflected by price only once tagging the rising 30-day moving trend is up), a six-day exponential moving average (EMA) and average). Finally, any thoughts that divergence between the a three-day RSI. (The choice of lookback period is a function of RSI peaks and the price peaks forewarn of a reversal should be the desired number of trades and the typical volatility of the discarded with such a short-term RSI lookback period. A simple guideline can be derived from this information: If market. For example, if you choose a 14-day RSI, you would the 30-day moving average is rising and the three-day RSI have far fewer signals compared to using a three-day RSI.) Each time the market closes below its six-day EMA (the bars drops below 50, go long on the next up close. This is a good indicated by A, C, E, G and I) the RSI moves below 50, which example of how a short-term RSI can be used to identify swing is the indicator’s median line. In other words, RSI readings trade opportunities in the upward zigzags of a rising trend. Why not just buy on the first up close following a close below 50 mean the market has closed below its six-day moving below the six-day EMA, which we’ve already defined as equivaverage. More interesting, however, is that the day after each of these alent to the three-day RSI? Because of the subtle but valuable bars, the market closed up and the RSI rose, sometimes cross- information provided by the level from which the RSI turns ing above 50 and sometimes not. Following bar B’s higher back up: During a rising market the RSI will bottom around 40. close, the market rallied for two more days; following bar D it (More about this in a moment.) rallied one more day; following bar F it rallied two more days; after bar H it rallied three more days; and following bar J the The RSI as trend indicator Figure 2 (above) shows a downtrending market with a threemarket moved up two more days. A couple of other points are worth noting. First, in February, day RSI and a 30-day moving average, but no six-day moving the RSI dropped only to between 40 and 50 and then rallied average, to give a clearer view of the price action. At bar A the near (or above) 75. Although this is a very short-term RSI, the 30-day moving average is falling and the RSI makes its first fact that the RSI turned back up from such high “oversold” lev- move above 50. The following day (bar B) the RSI declined and els was a sign of strength. Relative strength index values shift
ACTIVE TRADER • September 2003 • www.activetradermag.com 17

FUTURES

&

OPTIONS

Trading Strategies continued

FIGURE 3 RSI INDICATIONS The lower closes at point A and B established a support level in the RSI. When the indicator fell below that level at point C, it signaled a trend change.
Nasdaq 100 E-Mini (NQ), daily 1,150

1,100

1,050

1,000 30-day SMA 950

900 RSI 75 50 A November 1 Source: CQGNet 11 18 B 25 C December 2 25 9

FIGURE 4 SIGNALLING A TREND SHIFT The move at point C above the RSI resistance level indicated a switch from downtrend to uptrend.
Nasdaq 100 E-Mini (NQ), daily 1,100 30-day SMA 1,075

1,050

1,025

1,000

RSI A B C

D 75 50 25 13

E 9 Source: CQGNet 16 23 30 January 2003 2 6

the market closed lower that day and the next. The next day the market and RSI both closed higher, but still below their respective readings at bar A. The market then dropped for six days. At bar C, the RSI closed above 50; after bar D the market fell three days; and after bar F the market dropped one day. During this downtrend, the RSI peaked each time below 75 and dropped well below 25 on down swings because of the influence of the downtrend, which pushed the RSI to a lower range — the opposite of the situation in the Figure 1 uptrend. The typical trend-influenced shift in RSI readings can be used both to confirm the trend and to warn of a trend change. For example, in Figure 3 (left) the EMini Nasdaq 100 futures are in an uptrend (indicated by the rising 30-day moving average), and at points A and B the RSI turns back up from lows at 45. However, at point C the RSI falls to 35, closing below the support level established by points A and B. In other words, there was a downward shift in the trend that pushed the RSI readings lower well before the market crossed below the 30-day moving average. Figure 4 (left) is an example of an upward shift in RSI readings against the trend. In this case, the 30-day moving average is rolling over, indicating the market trend is turning down, which is also indicated by the RSI peaks (points A and B) that occur just below 60 and the following RSI bottom that is well below 25. But at point C, the RSI climbs as high as 75 and breaks the RSI resistance line, an early sign the trend was turning up. Next, the RSI moved well above 75, and at the first pullback (point D) the RSI bottomed at 50, more evidence of an upward trend. This behavior leads to a second guideline: Support lines drawn along RSI troughs during an uptrend should be just below 50. During a downtrend, resistance lines connecting RSI peaks should be just above 50. Violation of these trendlines warns the trend is changing. Also, consider RSI readings above 75 not as entry points, but as signs the market is strong or turning strong — i.e., confirmation of an uptrend. The opposite is true of RSI readings of 25 or lower confirming a downtrend.

18

www.activetradermag.com • September 2003 • ACTIVE TRADER

The relative strength index (RSI)
The relative strength index (RSI) is an oscillator that ranges from 0 to 100. The formula is: RSI = 100 - (100/[1+RS]) where RS = relative strength = the average of the up closes over the calculation period (e.g., 10 bars, 14 bars) divided by the average of the down closes over the calculation period. When calculating a 10-day RSI, if six days closed higher than the previous closes, subtract the previous close from the current close for these days, sum the differences, and divide the result by 10 to get the up-close average. (The sum is divided by the total number of days in the lookback period and not the number of up-closing days.) For the four days that closed lower than the previous day’s close, subtract the current close from the previous close, sum these differences, and divide by 10 to get the down-close average. If the up-close average was .8 and the down-close average was .4, the relative strength over this period would be 2. The resulting RSI would be 100 (100/[1+2]) = 100 - 33.3 = 66.67.

Exponential moving average (EMA)
The simple moving average (SMA) gives every price point in the average equal emphasis, or weight. Weighted moving averages emphasize more recent price action. The exponential moving average (EMA) weights prices using the following formula: EMA = SC * Price + (1 - SC) * EMA(yesterday) where SC = a “smoothing constant” between 0 and 1, and EMA(yesterday) = the previous day’s EMA value. You can approximate a particular SMA length for an EMA by using the following formula to calculate the equivalent smoothing constant: SC = 2/(n + 1) where n = the number of days in a SMA of approximately equivalent length. For example, a smoothing constant of .095 creates an exponential moving average equivalent to a 20-day SMA (2/(20 + 1) = .095). The larger n is, the smaller the constant, and the smaller the constant, the less impact the most recent price action will have on the EMA.

Further experimentation
The RSI was used here to find trade opportunities in the direction of the trend, as well as to indicate the trend was shifting direction. This is not the only tool that could function in this role — other oscillators could be used as well. The key is to review market action and determine practical criteria based on the typical behavior of the oscillator within the context of a trend. Once that behavior is quantified, you can develop entry strategies and risk management points for trades.

Trading approach highlights
Using the RSI to catch swing points 1. Define the trend: If the n-day (default: 30-day) simple moving average (SMA) is rising, the trend is up; if the SMA is falling, the trend is down. 2. Use a short-term RSI to find swing entry points: If the 30day moving average is rising and the three-day RSI drops below 50, go long on the next up close. Reverse for short trades. Using the RSI as a trend indicator 1. Support lines drawn along RSI troughs during an uptrend should be just below 50. During a downtrend, resistance lines connecting RSI peaks should be just above 50. Violation of these levels warns the trend is changing. 2. Consider RSI readings above 75 as signs the market is strong or turning strong — i.e., confirmation of an uptrend, not as entry points. The opposite is true of RSI readings of 25 or lower and downtrends.

Additional research
“Indicator Insight: Relative strength index (RSI),” Active Trader, August 2001, p. 88 “Indicator Insight: Weighted and exponential moving averages,” Active Trader, July 2002, p. 76 “The trend-following RSI,” Active Trader, January 2003, p. 44

ACTIVE TRADER • September 2003 • www.activetradermag.com

19

TRADING Strategies

RETRACEMENT TENDENCIES
Traders often calculate likely retracement levels, but
BY THOM HARTLE

is there any validity to such techniques? Objectively defining trends and retracements makes it possible to analyze the price swings in a market and determine

he market never moves in a if retracements follow any noticeable patterns. straight line — it repeatedly swings up and down on a shorter-term basis as it moves in the direction of the longer-term trend. FIGURE 1 DEFINING PRICE MOVES During a bull trend, the market advances and then retraces part of each Solid green lines indicate uptrending conditions and dashed red lines repre previous rise, repeating the process until sent downtrending conditions. the long-term trend turns down. Similarly, during a bear trend, the market zigzags A Nasdaq 100 index-tracking stock (QQQ), daily downward, with upswings retracing the 42.00 down legs of the prevailing trend. As such, retracements represent opportunities to get on board a trend. Traders often project retracement, or pullback lev40.00 C els to use as entry points. K E It’s easy to see retracements on a chart in hindsight. If a market advances or 38.00 G declines overall for six months, it’s obvious how each countertrend swing repreB D sented a retracement and potential entry 36.00 point. I To take advantage of retracement F opportunities as they develop, it’s neces34.00 sary to first have objective definitions of J both trends and retracements. Then, you H can determine if retracements display any noticeable patterns that could be used in 1/2/2002 2/1/2002 3/1/2002 the future. For example, if you could Source: Fibonacci Trader determine with a measure of certainty the next retracement in an uptrend would be Before analyzing percentage retracements, we must first 50 percent of the previous upswing, it would greatly simplify establish a definition for a trend. entering the market on the long side. Some traders believe the market frequently retraces in percentages derived from the numbers in the Fibonacci series (see Defining trends “The Fibonacci series”). The most common Fibonacci There are a number of ways to define trends, such as moving retracement targets corrections of 38.2 percent, 50 percent and averages and regression analysis. All trend-defining tech61.8 percent of the previous price move. niques have limitations, the primary one being lag. No matter To determine if there are such identifiable retracement per- what a trend indicator says, you can never be sure a trend has centages, we analyzed two years of daily Nasdaq 100 index- changed direction until after (sometimes well after) the fact. tracking stock (QQQ) prices from January 2002 to October We’ll use a price-based technique here. Let’s start by defin2003. The January 2002 opening price was $42.11. The market ing an uptrend as a pattern of consistently higher highs and fell to a low of $19.76 in October 2002 before climbing back to higher lows and a downtrend as a pattern of consistently lower $36 in October 2003. Because, the QQQ experienced both highs and lower lows. This means support levels will progresfalling and rising trends, we can be confident there is no sively climb (higher lows) during uptrends while resistance unusual bias to the data in that regard. levels will fall (lower highs) during downtrends.
20 www.activetradermag.com • March 2004 • ACTIVE TRADER

T

FIGURE 3 POST-RETRACEMENT MOVES FIGURE 2 RETRACEMENT SIZES There were 28 retracements, but there was no particular percentage size that appeared unusually often.
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Although two of the follow-through moves after retrace ment were noticeably large, there was no pattern to these price swings.
300% 250% 200% 150% 100% 50% 0% 1 2 3 4 5 6 7 8 9 10 1 11 21 31 41 51 6 1 71 81 9 2 02 12 2

1

3

5

7

9

11 13 15 17 19 21 23 25 27

If the market has been in a downtrend, a price breakout above a resistance level will be the sign the trend has turned up. If the market has been in an uptrend, a breakdown below a support level will be the signal the trend has turned down. This approach allows the immediate market action, rather than an arbitrary moving average lookback period, to determine the trend.

Defining support and resistance levels
It’s also necessary to objectively define support and resistance levels. The Gann swing rule, which provides precise rules for determining swing lows (support) and swing highs (resistance), is one possible approach. This technique uses consecutive highs or lows to determine the direction of the swing, as exemplified by the “three-day swing rule.” If the market makes three consecutive higher highs, the swing direction is up, and remains so until the market makes three consecutive lower lows. In this case, though, we will use the equivalent two-day swing rule. In his 1997 book, A W. D. Gann Treasure Discovered, Robert Krausz showed a copy of a page from one of Gann’s study courses that showed a hand-written note by Gann in the margin stating a two-day swing rule worked better than the three-day rule. “The two-day Gann swing charting technique” (p. 56) explains the definitions we will use. The trend is determined according to whether the market has broken above resistance (which indicates the trend is up) or below support (which means the trend is down). Once a trend is established we will measure the following three kinds of price moves: the size of each retracement as a percentage of the previous trend move; the size of the subsequent (postretracement) move back in the direction of the trend as a percentage of the preceding retracement swing; or, if the market fails to follow through in the direction of the previous swing, the size of the retracement before a trend reversal. For example, Figure 1 (opposite page) shows the QQQ price action during the first three months of 2002. Solid green lines indicate an uptrend and dashed red lines indicate a downtrend, based on the two-day Gann swing rule. For example, in mid-January, the line turned from solid green to dashed red because the market closed below the early January support
ACTIVE TRADER • March 2004 • www.activetradermag.com

level, changing the trend from up to down. The price swings can all be defined relative to the moves that preceded them. Starting with the early January peak (point A), the first downswing was a drop from $42.60 to $37.33 (B), a $5.27 decline. The market then rallied to $39.40 within the downtrend (C), which constitutes a 39-percent retracement of the A-B decline. Next, the market fell to $36.85 (D), a move that was 123-percent the size of the B-C retracement. Moving forward, the D-E swing retraced 80 percent of the BC swing and the subsequent E-F swing was 193 percent of the DE. The F-G upswing retraced 87 percent of the E-F downswing, and the drop from G to H was 172 percent of the F-G move. FIGURE 4 FAILED PRICE SWINGS Post-retracement follow-through swings that failed were as inconsistent as the other types of price swings.
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

1

2

3

4

5

6

The market then rose to point I, which marked a 54-percent retracement of the previous downswing. After retracing 73 percent of the H-I move (at J), the market rallied and closed above the resistance level established by the point I swing high, turning the trend from down to up. The swing from J to K was 309 percent of the I-J swing.

Retracements
If a market trended upward in stair-stepping fashion making 77 total up moves without ever violating a previous support level, there would be 38 retracements of those up moves.

21

FIGURE 5 REVERSING A REVERSAL The trend changed from up to down when price fell below the point-A support level, but it turned back up when price rallied above point B.
Nasdaq 100 index-tracking stock (QQQ), daily

D
30.40 30.00 29.60

ments that preceded a trend reversal. The six failed swings ranged from just under 30 percent to just over 80 percent. In short, none of the statistics here are consistent.

Trend reversals

The market changed trend by breaking through a previous support or resistance 28.80 level 20 times over the test period. Because the first trend change occurred in 28.40 early January and was based on a support level outside the study period, it will be 28.00 ignored, leaving 19 reversals to analyze. Six of these remaining reversals were A 27.60 preceded by the failed swings shown in C Figure 4. Interestingly, of the six times the 27.20 trend changed direction after a failed swing, it reversed again immediately in 5/5/2003 5/12/2003 5/19/2003 5/27/2003 6/2/2003 the direction of the prior trend four times. Source: Fibonacci Trader When the market changed trend by breaking support or resistance, it was Our two-year study of the QQQ contained 76 trending price from a greater than 100-percent price move of the previous swings, but only 28 retracements — nearly 25-percent fewer swing 13 times. In other words, if the market was in an than expected. The remaining swings were trend reversals, uptrend, it reversed and took out a previous support level by dropping virtually straight to it. most of which were larger than 100-percent countermoves, relIn Figure 5 (above), for example, swing A is a downswing ative to the previous swing. Figure 2 sorts the 28 retracements from smallest to retracement of a previous upswing, and swing B is a move to a largest. Notice there is no tendency — no cluster of readings at new high. But swing B-C retraces all of swing A-B, and by breaka certain percentage, which means that commonly referenced ing the swing low support level at point A, the trend turns down Fibonacci retracement levels did not make their presence felt — but only briefly. The market soon reverses again and rallies past the swing-high resistance level (B), changing the trend to up during the test period. (There is precisely one 50-percent again. In this case, there were two back-to-back trend reversals. retracement.) Figure 6 (below) shows the percentage moves relative to the Now let’s look at the price swings that followed the retracements — the post-retracements price swings back in the direc- previous swing for the 19 trend reversals, including those that tion of the prevailing trend. There were 22 such “follow- were percentage comparisons to a failure swing. For example, from Figure 5, the B-C downswing was 128 percent of A-B (bar through” moves after completed retracements (the remaining six were failed swings, which will be addressed separately). Figure 3 displays the size of these moves as a percentFIGURE 6 TREND REVERSAL age of the preceding retracement move, and sorts them from smallest to largest. The observations that are just barely larger Here, price swings that reversed the trend are sorted than 100 percent (far left) indicate the market followed through from low to high in terms of their size relative to the past the beginning of the retracement, only to reverse again previous swing. and either retrace a portion of the swing or the entire swing (and more), signaling a reversal. 350% As was the case with the retracements, there does not seem 300% to be any particularly consistent or typical percentage size to the follow-through moves. The final two percentages are high250% er than expected, but two out of 22 does not indicate anything 200% significant. 150% With 22 of the 28 retracements having greater than 100-percent follow through, the remaining six were failed price swings 100% that were themselves reversed — for example, price swing I-J 50% in Figure 1, when the market failed to make a new low. When 0% the market subsequently rallied above point J, the trend 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 switched from down to up. Figure 4 shows the percentage sizes of the six retraceB
29.20 22 www.activetradermag.com • March 2004 • ACTIVE TRADER

The two-day Gann swing charting technique
7 in Figure 6), and upswing C-D was 247 percent of B-C (bar 15 in Figure 6). Referring again to Figure 1, the J-K swing was 309 percent of the I-J swing, which is also the final observation in Figure 6. The interesting thing in Figure 6 is the larger higher-percentage swings to the right of center compared to Figure 3, which showed the sizes of the postretracement follow-through moves in the direction of the trend. Of the 19 trend reversals, the market continued in the direction indicated by the break of support or resistance trend level only five times. Five other times the market immediately reversed again, as in Figure 5. In other words, there were 10 times the market made back-to-back swing moves violating a previous support or resistance point, signalling a trend reversal, without an interim swing move in the opposite direction. One time the market made four consecutive backto-back violations of the previous support and resistance (think of the broadening triangle chart pattern). two-day swing consists of two consecutive higher highs or lower lows. In Figure 7, the first two consecutive higher highs are labeled 1 and 2 on the left side of the chart. This makes the swing direction up, and the green line indicates there was a previous breakout above resistance (not shown), which is the definition of an uptrend (see main story). The market rises, and then makes two consecutive lower lows, down to point A. This is a downswing in an uptrend. Following point A, the market FIGURE 7 TWO-DAY SWINGS makes two consecutive higher highs, leading to new highs. This is 2 E an upswing in an uptrend. Point A is 1 now a swing low and a support C 2 point. 12 Next, the market drops below 1 the support point A, established by the downswing and upswing. This 1 changes the trend to down at point 2 2 B, and the Gann Swing line becomes 1 A B a dashed red line. 2 1 1 The market then makes two consecutive higher highs to point C. 2 This is an upswing in a downtrend. D The two consecutive lower lows down to point D means this is a downswing in a downtrend. Point C is a swing high or a resistance point. The market reverses and makes two consecutive higher highs (an upswing in a downtrend) until it breaks the resistance established at point C. Because this was a downtrend, the trend now changes to up. Finally, the market makes two Source: Fibonacci Trader consecutive lower lows, which is a downswing in an uptrend.

A

Overcoming subjectivity

The difficulty of performing this type of analysis begins with the challenge of developing objective criteria for trends, trend reversals and retracements. Some nuance is inevitably lost when creating a framework that can accommodate all price action. As a result, some might argue that a trend reversal indicated by the method used in this study is not (to them, based on their own interpretation of the chart) really a reversal and, therefore, a retracement is still unfolding, and so on. However, without clear-cut definitions, you are left with ambiguity, rather than numbers on which to gauge probabilities and build trading procedures. Having said this, the study provided some interesting insights about retracements. First, there was no consistent percentage retracement size of prior price swings — Fibonacci or otherwise. Similarly, if a retracement held, and the market did continue in the direction of the trend, the market gave no evidence of a consistent percentage size for these moves, either. Also noteworthy was the number of times the market broke a support or resistance level, only to immediately reverse with substantial follow-through movement, as shown in Figure 6. This indicates the need for procedures to take advantage of such false breakouts. One name for these kinds of patterns is spring (false break of support) and upthrust (false break of resistance).

Don’t count your retracement level before it’s touched
The market does have periods when it trends in a progresACTIVE TRADER • March 2004 • www.activetradermag.com

sive, stair-stepping fashion, either up or down, as shown in Figure 1. However, it also can trade in a manner that would lead one to conclude that support and resistance levels, like records, were made to be broken. In this sense they represent targets for the market to flush out traders with stops just beyond these levels — which is exactly how markets function. Why is this the case? The collection of individuals in the marketplace probably does not change that much on a day-today basis, and if the market is going to go somewhere, it has to change the general sentiment to set the stage for the next price swing. If everyone is very bullish, all the available capital has probably been put to work. If this is the case, there is no one left to buy. This is why the market ebbs and flows, and people get in and out of positions, either by design or pain. But a break of support or resistance can lead to a positive or negative sentiment extreme, creating a situation in which a real opportunity exists for a big trade.Ý

23

FUTURES

&

OPTIONS

Trading Strategies

Percent b
Putting a new twist on an old indicator results in a tool that determines short-term market bias and sets up intraday swing trades — with a little help from a price pattern.
BY THOM HARTLE

T

he eternal quest for would-be short-term traders sents the oversold threshold and the moving average repreis to find a way to identify swing points consis- sents the median line. The strategy discussed here uses the percent b indicator both tently and mechanically. It cannot be done, but traders never tire of experimenting with this or to define the prevailing market bias and, combined with a that indicator in the hope it will lead to a universal rule for three-bar price pattern, to signal trade entries in the E-Mini identifying short-term exhaustion points, or overbought and S&P 500 futures (ES) on an intraday basis. oversold levels. Oscillators are the indicators typically enlisted for this role, and although there are literally dozens of tools that fall FIGURE 1 PERCENT B under this heading, virtually all of them The percent b indicator shows where the closing price is relative to the rely on similar momentum calculations upper and lower Bollinger Bands (indicator levels 100 and zero, respectively). that measure how much price has moved
Continuous S&P 500 E-Mini futures (ES), five-minute 1,074 Upper Bollinger Band 1,072

up or down over a recent time period. Typically, a high indicator threshold defines prices that are theoretically overbought and a low indicator threshold identifies prices that are theoretically oversold. A middle equilibrium or median line separates rising and falling momentum conditions. Bollinger Bands are a similar tool that use a statistical calculation to measure the degree to which price has moved above or below its average. The indicator consists of lines placed (by default) two standard deviations above and below a moving average of closing prices (see Figure 1, right). For a more detailed discussion of Bollinger Bands, see “Bollinger Bands, variance, and standard deviation,” and the references listed at the end of the article. The percent b indicator, which was developed by John Bollinger, measures the closing price relative to the upper and lower Bollinger Bands. In a way, the indicator is a more typical oscillator-type depiction of Bollinger Bands for which the upper band is equivalent to the overbought threshold, the lower band repre24

1,070

1,068 20-bar simple moving average Lower Bollinger Band Percent b 1,066 150 100 50 0 Volume ˇ 16-8:30 Source: CQGNet 9:30 10:30 11:30 12:30 13:30 16-14:30
15,000 10,000 5,000 0

www.activetradermag.com • April 2004 • ACTIVE TRADER

Percent b basics
The percent b indicator reflects closing price as a percentage of the lower and upper Bollinger Bands according to the following formula: Percent b = [(Close – lower Bollinger Band)/(Upper Bollinger Band – lower Bollinger Band)]*100

above or below any of the lines. This is the information percent b provides.

Highlighting the market bias

To the untrained trader, a price move beyond a Bollinger Band (upper or lower) implies a market is at an extreme, having moved more than two standard deviations from the 20-bar average price. The natural conclusion is the market is overIf the closing price is the same as the upper Bollinger Band bought or oversold and should reverse. value, percent b would be 100 (percent), and if the closing price Although this is often true in the very near term, the fact a is above the upper Bollinger Band, percent b would be greater market has moved to an extreme level is, first and foremost, a than 100. If the close is equal to the moving average, percent b sign of momentum strength — a trend. Therein lies a trading opportunity. In Figure 1, the percent b indicator FIGURE 2 DOWNTRENDING MARKET BIAS closed above 100 (meaning price closed above the upper Bollinger Band) a total The fact that percent b falls below zero and never trades above 100 means of five times and never closed below zero the market bias is down. (i.e., price did not close below the lower Bollinger Band). The first percent b close Continuous S&P 500 E-Mini futures (ES), five-minute above 100 occurred on the first five1,085.0 minute bar (8:30 a.m. CT) of the first full Upper Bollinger Band 1,082.5 trading day on the chart. 20-bar simple moving average The market dropped to the lower 1,080.0 Bollinger Band, but did not close below it, which means percent b held above 1,077.5 zero. Throughout the trading day percent b sometimes dipped below 50 but 1,075.0 stayed above zero while the market trended higher for the day. Overall, per1,072.5 Lower Bollinger Band cent b was indicating the market had an 1,070.0 upside bias. Figure 2 shows the opposite circum1,067.5 stance. In this case, percent b broke below zero on the first bar and subsequently pushed below zero (that is, price 100 Percent b breaks zero No values above 100 dropped below the lower Bollinger Band) a few more times. It rallied above 50 50 but never as high as 100. In this case, percent b indicated the market bias was 0 down.
-50 Volume ˇ 15-8:30 9:30 Source: CQGNet 10:30 11:30 12:30 13:30 15-14:30
20,000 10,000 0

Setting up trades

is 50 percent, and if the close is equal to the lower Bollinger Band, percent b would be zero. If the close is below the lower band, percent b would be negative. Figure 1 shows these relationships on a five-minute chart of the E-Mini S&P futures using 20-bar Bollinger Bands and percent b. Although Bollinger Bands show you where price is relative to its three lines, it does not tell you the degree to which it is
ACTIVE TRADER • April 2004 • www.activetradermag.com

The percent b indicator is not used to trigger trades, but rather to signal favorable conditions for a trade. If percent b indicates the market bias is up, look for pullbacks to enter long trades and take profits as the uptrend reasserts itself. Returning to Figure 1, notice the market traded back below the 20-bar simple moving average a number of times. Many people believe a price drop below a moving average signals a downtrend. But if you have a tool that defines the market bias to be up, then a price move below the average is more appropriately interpreted as a buying opportunity because the mar25

FUTURES

&

OPTIONS

Trading Strategies continued

Bollinger Band, variance and standard deviation
Bollinger Bands consists of three lines, as shown in Figure 1 (p. 66): Upper band = 20-period simple moving average + 2 standard deviations Middle line = 20-period simple moving average of closing prices Lower band = 20-period simple moving average - 2 standard deviations The upper and lower bands represent levels at which price is trading higher or lower, respectively, than it has approximately 89 percent of the time over the past 20 days. Standard deviation measures the degree to which values in a group (say, closing prices from day to day) vary from the mean value. Standard deviation is derived from the variance of a set of numbers which, mathematically, is the average squared “deviation” (or difference) of each number in the group from the group’s mean value, divided by the number of elements in the group. For example, for the numbers 8, 9 and 10, the mean value is 9 and the variance is: {(8-9)2 + (9-9)2 + (10-9)2}/3 = (1 + 0 + 1)/3 = 0.67 Now look at the variance of a more widely distributed set of numbers — 2, 9, 16: {(2-9)2 + (9-9)2 + (16-9)2}/3 = (49 + 0 + 49)/3 = 32.67 The second group of numbers are more widely distributed and thus has a higher variance. Standard deviation is simply the square root of the variance. So, the standard deviation of 8, 9 and 10 is: 0.67 = 0.82 The higher the standard deviation, the greater the volatility. The proper use of standard deviation calls for a sample of at least 30 observations, in which case, statistically, 95 percent of values will fall within two standard deviations of the mean value. Because Bollinger Bands typically use only 20 days of closing prices, the bands will generally encompass 88 to 89 percent of the price action.

ket has become cheap relative to the market bias, or trend. The percent b readings in Figure 2 indicated the bias or trend was down. In this case, prices above the 20-bar simple moving average meant the market had become expensive relative to the bias. Shrewd traders seek to buy when price is cheap in an uptrending market and sell short in a downtrending market when price is relatively expensive. Next, we’ll illustrate a pattern that can be used to place trades when these conditions are in effect.

Trade entry
An extreme percent b reading dictates the market bias or trend. Although this may run counter to traditional interpretations of oscillator-based approaches (in that extreme readings are usually considered signs of an overbought or oversold market, and an indication of an impending reversal), careful scrutiny reveals that during trends, oscillator readings shift. That is, they have persistently higher high and higher low readings in uptrends, and lower low and lower high readings in downtrends. With this in mind, we’ll trade based on the following concept: If the trend is up, we’ll look for a buy (price) setup once price trades back below the 20-bar simple moving average — in other words, when percent b moves below 50. In this case, we’ll use the simple “pivot low” pattern shown in Figure 3 (p. 70): a three-bar pattern in which bar 1 is followed by a bar with a lower high and lower low (bar 2), followed by a bar with a higher high and higher low (bar 3). Figure 3 depicts an ideal setup. The percent b indicator traded above 100, signaling an uptrend. Next, percent b dropped below 50, indicating price was cheap on a relative basis. At this point, you would watch for a pivot low and confirm percent b has not dropped below zero. Go long when price moves above bar 2’s high. If the market is strong, this pivot low (the low of bar 2) should act as a support level and should not be taken out. Therefore, an appropriate risk point is just below the low of bar 2. Profits can be taken based on an evaluation of recent volatility (i.e., how much the market has tended to move lately), a

Additional Active Trader reading
“Indicator Insight: Bollinger Bands,” July 2003, p. 74. “Relatively speaking: John Bollinger,” April 2003, p. 60. “John Bollinger: Focus on the markets,” January 2001, p. 74. Past articles can be purchased and downloaded through the Active Trader online store at www.activetradermag. com/purchase_articles.htm.

26

www.activetradermag.com • April 2004 • ACTIVE TRADER

FUTURES

Trading Strategies
continued

&

OPTIONS

FIGURE 3 THE PIVOT PART OF THE EQUATION The percent b indicator signals conditions are favorable for an up move; the three-bar pivot triggers the actual trade.
Continuous S&P 500 E-Mini futures (ES), five-minute 1,076 Upper Bollinger Band 1,075 1,074 1,073 20-bar simple moving average 1,072 1,071 1,070 Lower Bollinger Band Percent b breaks 100 100 50 Percent b drops below 50 Volume 0
15,000 10,000 5,000 0

trailing stop, or a retest of the high of the day. (It is always important to track recent volatility when taking profits, as this is not a stable attribute in the markets.) Figure 4 shows an ideal short selling opportunity. The percent b dropped below zero around 12:15 p.m., signaling the trend was down. Next, the percent b rose above 50, indicating price was expensive relative to the trend. As long as percent b has not crossed above 100, look for a pivot high. Go short when price drops below the low of bar 2. And if the trend is down, this resistance level should hold, so the risk point should be just above the high of bar 2. Again, for taking profits, you can use a retest of the previous support point, such as the low for the day, a trailing stop or a preset target based on the current volatility.

1 2

3

13:30 Source: CQGNet

13:45

14:00

14:15

14:30

14:45

17-15:00

Potential weaknesses
It is always wise to consider the circumstances that would cause a trading approach to fail. In the case of a percent b setup, a market may not have a true bias during a congestion period, so there will be periods when support and resistance levels established by the pivot lows and highs will be violated rather than hold. Also, the move by percent b back to the 50 level might be the first leg in a trend reversal rather than an opportunity to trade in the direction of the trend. Unfortunately, neither situation is easy to ascertain until after the fact.

FIGURE 4 SHORT TRADE In this case, the percent b reading below zero indicated a downtrending market and the move above 50 signaled a pullback or correction had occurred that set up a trade. The sell is executed when price drops below the low of bar 2.
Continuous S&P 500 E-Mini futures (ES), five-minute

2 1

1,073

3
Upper Bollinger Band 1,072 1,071 1,070

Turning a concept inside out
Trading is the process of converting an observable market attribute into rules and procedures that capitalize on price moves. The oscillator-based percent b approach is based on the observation that high or low indicator readings actually portend further directional movement, not the more commonly expected reversal. Combining this view with a simple price pattern offers a basic trading approach that can be modified and augmented.

Lower Bollinger Band

20-bar simple moving average

1,069 1,068

Percent b rises above 50 Percent b breaks zero 100 50 0 Volume
15,000 10,000 5,000 0

12:00 Source: CQGNet 27

12:15

12:30

12:45

13:00

13:15

3-13:30

www.activetradermag.com • April 2004 • ACTIVE TRADER

TRADING Strategies

Quarterly momentum and ETFs
In certain situations, the market tips its hand regarding which sectors are likely to stay atop the market. Certain signals early in each quarter can help locate sectors that are likely to remain strong through the end of the quarter.

FOLLOWING THE MONEY :

BY THOM HARTLE

xchange Traded Funds (ETFs) — a group of related equity products that represent different indices, market sectors and groups, and which trade like stocks — have been enormously successful since the first few were launched in the 1990s. The two most popular are the Nasdaq 100 index-tracking stock (QQQ) and S&P 500 indextracking stock (SPY). However, there are many more available — more than 120 are listed at the Nasdaq Web site for evaluation (see “Funds for all,” Active Trader, November 2002 ), ranging from sector indices (biotechnology, IBB) to country-specific instruments (the Australian stock market, AAA) to Treasury bond index funds (TLT).
28

E

This abundance of ETFs enables traders who expect a certain investment “theme” to become popular on Wall Street to position themselves accordingly.A working hypothesis of some traders is that when an influential investment house starts buying certain sectors or market areas, other big money will flow in that direction for fear of being left behind. Fund managers will often make big allocation decisions on a quarterly

basis. So, for example, traders who believe money will move into semiconductor stocks could buy the Merrill Lynch Semiconductor HOLDR (SMH), an ETF that represents a basket of 20 semiconductor stocks. This approach is predicated on the ability to spot the emergence of a new investment theme. One way to approach this is to monitor the initial trading each quarter with the expectation the emerg-

www.activetradermag.com • June 2004 • ACTIVE TRADER

Top 50 ETFs
Name Dow Jones index-tracking stock Fresco DJ EURO STOXX 50 iShares Dow Jones U.S. Real Estate index fund iShares Lehman 1- to 3-year Treasury bond fund iShares Lehman 7- to 10-year Treasury bond fund iShares Lehman 20-year Treasury bond fund iShares MSCI - Austria iShares MSCI - Brazil iShares MSCI - EAFE iShares MSCI - EMU index fund (European Monetary Union) iShares MSCI - Hong Kong iShares MSCI - Japan iShares MSCI - Malaysia (Free) iShares MSCI - South Korea iShares MSCI - Taiwan iShares MSCI - United Kingdom iShares MSCI Emerging Markets iShares Nasdaq Biotechnology index fund iShares S&P 500/BARRA growth index fund iShares S&P 500/BARRA value index fund iShares S&P 500 index fund iShares Russell 1000 growth index fund iShares Russell 1000 index fund iShares Russell 1000 value index fund iShares Russell 2000 growth index fund iShares Russell 2000 index fund iShares Russell 3000 index fund Merrill Lynch B2B Internet HOLDRS Merrill Lynch Biotech HOLDRS Merrill Lynch Internet Infrastructure HOLDRS Merrill Lynch Internet HOLDRS Merrill Lynch Market Oil Service HOLDRS Merrill Lynch Pharmaceutical HOLDRS Merrill Lynch Retail HOLDRS Merrill Lynch Semiconductor HOLDRS Merrill Lynch Software HOLDRS Merrill Lynch Utilities HOLDRS MidCap SPDR Trust Series I Nasdaq-100 index-tracking stock Select sector SPDR fund - Basic Industries Select sector SPDR fund - Consumer Discretionary Select sector SPDR fund - Consumer Staples Select sector SPDR fund - Energy Select Sector Select sector SPDR fund - Financial Select sector SPDR fund - Health Care Select sector SPDR fund - Industrial Select sector SPDR fund - Technology Select sector SPDR fund - Utilities S&P 500 Index-tracking stock Vanguard Total Stock Market VIPERs
Source: Nasdaq ACTIVE TRADER • June 2004 • www.activetradermag.com 29

Symbol DIA FEZ IYR SHY IEF TLT EWO EWZ EFA EZU EWH EWJ EWM EWY EWT EWU EEM IBB IVW IVE IVV IWF IWB IWD IWO IWM IWV BHH BBH IIH HHH OIH PPH RTH SMH SWH UTH MDY QQQ XLB XLY XLP XLE XLF XLV XLI XLK XLU SPY VTI

ing leaders will hold their advantage for the remainder of the quarter. (This is similar to an idea discussed by trader Linda Raschke regarding the subsequent strong performance of certain stocks that post the biggest gains at the beginning of a quarter. See “Linda Raschke keeps up the pace,” Active Trader, February, 2004). The following analysis investigates whether early-quarter momentum in individual ETFs tends to sustain itself, and how well early leaders compare to the performance of the QQQs and the SPYs.

Following the leaders
To see if the ETFs that show strength early in a quarter tend to keep outperforming the market over time, 50 of the currently most popular ETFs (based on average daily volume greater than 100,000 as of the end of 2003) were studied on a quarterly basis from the first quarter of 2001 through the fourth quarter of 2003. (Note: Only 40 of the current collection of ETFs meeting this criteria existed in the first quarter of 2001.) The leading ETFs at the beginning of a given quarter were defined as those with the largest percentage gains from the close of the last day of the previous quarter to the close of the second Friday of the current quarter. For example, the initial two-week performance for the first quarter of 2001 was calculated by comparing the Jan. 12, 2001, ETF closing prices to the Dec. 29, 2000, closing prices. (Incorporating the second Friday of the month insured the monthly employment numbers, which are potentially market-moving statistics — see “Trading the monthly jobs report” — were reflected in the analysis.) The end-of-quarter performance for each of the top-five performing ETFs was then calculated. This was done by comparing the closing prices on the last day of the current quarter to the last day

TABLE 1 PERFORMANCE OF EARLY-QUARTER ETF LEADERS, 2001 Although the beginning of the first quarter of 2001 was very bullish, all but one of the ETFs — the iShares MSCT-Taiwan Index Fund (EWT) — were in the red by the end of the quarter.
First quarter First two weeks ETF HHH EWY SMH EWT XLK QQQ SPY %Bullish 19.68% 18.37% 13.16% 13.02% 11.88% 7.47% 0.62% 42.00% 1 2 3 4 5 9 18 End of quarter -9.81% -0.08% -15.33% 13.89% -20.80% -32.93% -11.05% 19 4 30 1 34 36 23 Second quarter First two weeks SWH 17.34% SMH 11.71% QQQ 9.32% IIH XLK XLB SPY 8.22% 7.02% 6.91% 1.85% 66.00% 1 2 3 4 5 6 23 End of quarter 35.91% 16.17% 16.73% 15.35% 13.59% 10.47% 5.06% 1 7 4 9 11 13 26 Third quarter First two weeks EWM 4.65% XLY UTH XLU XLB SPY 3.58% 2.58% 1.54% 1.50% -0.29% 16.00% 1 2 3 4 5 12 34 End of quarter 1.40% -16.99% -11.31% -6.07% -12.96% -14.81% -36.59% 2 22 8 4 9 13 37 BHH SMH IIH Fourth quarter First two weeks 34.63% 27.67% 20.38% 1 2 3 4 5 9 18 End of quarter 99.61% 41.58% 45.99% 18.11% 34.27% 26.86% 9.44% 1 5 4 18 7 9 32

1/12/01 Rank 3/30/01 Rank

ETF 4/12/01 Rank 6/29/01 Rank

ETF 7/13/01 Rank 9/28/01 Rank

ETF 10/12/01 Rank 12/31/01 Rank

SWH 21.61% QQQ 19.22% HHH 14.68% SPY 4.84% 84.00%

QQQ -5.58%

Source: data - eSignal; table - Excel

TABLE 2 PERFORMANCE OF EARLY-QUARTER ETF LEADERS, 2002 The fourth quarter was the only one in 2002 that had more than 50-percent bullishness among all ETFs. Three of the top-five ETFs at the beginning of the quarter were still near the top at the end of the quarter.
First quarter First two weeks ETF IIH SMH SWH BHH QQQ XLK SPY %Bullish 13.90% 8.24% 7.85% 7.21% 4.99% 4.58% 0.56% 31.00% 1 2 3 4 5 6 15 End of quarter -21.09% 11.18% -7.83% -20.27% -7.32% -9.71% 0.19% 45 5 39 44 38 41 28 IYR EWZ Second quarter First two weeks 4.92% 3.93% 1 2 3 4 5 27 38 End of quarter 2.35% -26.68% -3.38% 10.64% -9.44% -13.59% -27.62% 3 38 7 1 16 27 40 EWY EWT Third quarter First two weeks 6.53% 1.16% 1 2 3 4 5 9 22 End of quarter -20.01% -11.89% -26.23% -36.61% -40.31% -20.61% -17.35% 27 7 38 41 43 30 15 Fourth quarter First two weeks SWH 11.21% HHH XLK QQQ BBH SMH SPY 9.69% 9.55% 7.24% 6.20% 5.82% 2.90% 52.00% 1 2 3 4 5 6 16 End of quarter 29.26% 28.45% 25.11% 17.62% 8.55% 15.12% 7.87% 3 4 5 7 15 8 19

1/11/02 Rank 3/28/02 Rank

ETF 4/12/02 Rank 6/28/02 Rank

ETF 7/12/02 Rank 9/30/02 Rank EWM 3.32% SMH -2.11% EWZ -2.16% QQQ -4.98% SPY -7.18% 0.07%

ETF 10/11/02 Rank 12/31/02 Rank

EWM 3.89% EWO 2.78% IWM SPY 2.40% -2.71% 15.50%

QQQ -7.04%

Source: data - eSignal; table - Excel

TABLE 3 PERFORMANCE OF EARLY-QUARTER ETF LEADERS, 2003 Nearly 82 percent of the ETFs had positive returns in the first two weeks of the first quarter of 2003. This kind of broad-based strength was more likely to result in the early-quarter leaders maintaining their dominance through the rest of the quarter.
First quarter First two weeks ETF IIH BHH SWH SMH HHH QQQ SPY %Bullish 19.33% 16.23% 15.34% 15.30% 13.02% 11.20% 5.47% 81.60% 1 2 3 4 5 8 14 End of quarter -5.04% 17.28% -5.85% 4.24% 17.65% 3.61% -3.96% 34 2 38 5 1 7 28 Second quarter First two weeks EWY 10.85% FEZ EZU EWZ XLF SPY 9.72% 8.39% 6.23% 5.64% 2.84% 71.40% 1 2 3 4 5 17 35 End of quarter 33.09% 26.41% 26.91% 27.07% 18.26% 15.21% 18.61% 3 8 7 6 22 32 20 Third quarter First two weeks SMH 10.45% EWT 10.29% IBB EWY EWJ SPY 8.75% 8.10% 7.84% 2.67% 72.00% 1 2 3 4 5 10 28 End of quarter 21.86% 22.93% 7.11% 6.80% 20.50% 8.25% 2.38% 3 2 15 17 4 11 33 Fourth quarter First two weeks HHH 14.78% EWY 12.18% EWZ 11.98% EWM 11.71% SMH QQQ SPY 11.48% 7.99% 4.62% 88.00% 1 2 3 4 5 9 27 End of quarter 18.09% 17.28% 33.96% 9.51% 20.61% 12.46% 11.34% 8 9 1 33 6 20 27

1/10/03 Rank 3/31/03 Rank

ETF 4/11/03 Rank 6/30/03 Rank

ETF 7/11/03 Rank 9/30/03 Rank

ETF 10/10/03 Rank 12/31/03 Rank

QQQ 6.31%

QQQ 1.03%

Source: data - eSignal; table - Excel 30 www.activetradermag.com • June 2004 • ACTIVE TRADER

FIGURE 1 BEGINNING VS. END OF FIRST QUARTER, 2001 This ranking shows only a relative handful of the funds had strong returns (green bars) in the first two weeks of the quarter; the remaining funds post ed near breakeven to negative returns. This narrowness in the leadership early in the quarter belied the behind-the-scenes weakness in the market. When the leaders finally gave up, the entire market tumbled.
30.00% 20.00% 10.00% 0.00% -10.00% -20.00% -30.00% -40.00% -50.00% -60.00% -70.00% -80.00% Jan. 12, 2001, return

of the previous quarter. For instance, performance for the first quarter was calculated by comparing the March 30, 2001, closing prices to the Dec. 29, 2000, closing prices. Finally, the percentage returns from the top-five performers were compared to those of the QQQ and the SPY. Tables 1 through 3 show the quarterly returns for 2001, 2002 and 2003, respectively. The five ETFs that performed best in the first two weeks of each quarter are shown, along with the final quarterly returns and where each ETF ranked at the end of the month. The results for SPY and QQQ are included for each period as well. In Table 1, although the first several trading days of 2001 were spectacular (the top two ETFs were up nearly 20 percent), the market quickly turned down and all but one of the ETFs — the iShares MSCT-Taiwan Index Fund (EWT) — posted losses for the quarter. Figure 1 provides some explanation for this. Ranking all 40 ETFs by their performance in the first two weeks of 2001’s first quarter, along with their respective end-of-quarter 2001 returns, reveals only a relative handful of the funds had strong returns in the first two weeks of the quarter; the remaining funds posted near breakeven to negative returns. This narrowness in the leadership during the first two weeks of January indicated there was a great deal of weakness behind the scenes in the market and, as a result, when the leaders finally gave up, the entire market tumbled. To reflect the extent of the leadership in a given quarter, Tables 1 through 3 also include (in the last row, labeled “% Bullish”) the percentage of ETFs that had positive returns in the first two weeks of the quarter. The tables also show the QQQ and the SPY were below the top five the majority of the time — not surprising, since QQQ tracks 100 stocks and SPY tracks 500. Although this diversification spreads risk, it also lowers return. Also, what the QQQ and the SPY did in the first two weeks during the second,

March 30, 2001, return

Source: data - eSignal; chart - Excel

FIGURE 2 BEGINNING VS. END OF FIRST QUARTER, 2002 As was the case in 2001, only a handful of ETFs were positive in the first two weeks of the first quarter of 2002, while the rest were flat to down; just 30 percent had positive returns.
40.00% 30.00% 20.00% 10.00% 0.00% -10.00% -20.00% -30.00% Source: data - eSignal; chart - Excel March 28, 2002, return Jan. 11, 2002, return

FIGURE 3 BEGINNING VS. END OF FIRST QUARTER, 2003 Early in the first quarter of 2003, buying was strong across the board. The final first-quarter 2003 returns were mixed, but the action that occurred in the first two weeks of the period set the tone for the rest of the year.
25.00% 20.00% 15.00% 10.00% 5.00% 0.00% -5.00% -10.00% -15.00% -20.00% -25.00% Source: data - eSignal; chart - Excel Jan. 10, 2003, returns March 31, 2003, returns

ACTIVE TRADER • June 2004 • www.activetradermag.com

31

FIGURE 4 BROAD-BASED BUYING Demand was high in the first two weeks of 2003’s second quarter. Only a handful of ETFs had negative returns, and even those ended up positive by the end of the quarter.
45.00% 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% -5.00% April 11, 2003, returns June 30, 2003, returns

Source: data - eSignal; chart - Excel

FIGURE 5 4TH QUARTER 2001 The Merrill Lynch Internet HOLDR (BHH) was a steady climber in the fourth quarter of 2001. The remaining ETFs followed suit, but at a slower rate of ascent.
120.00% 100.00% 80.00% 60.00% 40.00% 20.00% 0.00% -20.00% BHH HHH IIH SMH SWH SPY QQQ

Source: data - eSignal; chart - Excel

FIGURE 6 4TH QUARTER 2002 In the fourth quarter of 2002 the ETFs peaked in early December and then declined. BHH again led the ETF pack.
0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 -0.2 BHH HHH SMH SPY SWH QQQ XLK

Source: data - eSignal; chart - Excel 32

third and fourth quarters of each year indicated the outcome for the top-five ETFs: Positive QQQ and SPY returns correlated to positive ETF returns and negative returns also were hand-inhand. Figure 2 is similar to Figure 1 in that it shows only a handful of ETFs were positive in the first two weeks of the first quarter of 2002, while the rest were flat to down; just 31 percent had returns greater than zero. Figure 3, on the other hand, shows how broad based the buying was in the first two weeks of January. Although the final first quarter 2003 returns were mixed, most of the action occurred in the first two weeks of the period, and this strength set the stage for the rest of the year. Table 3 shows nearly 82 percent of the group had positive returns in the first two weeks. Figure 4 shows how demand was high in the first two weeks of 2003’s second quarter, as there was just a handful of ETFs with negative returns during this period, and even those ended up positive by the end of the quarter. An important aspect of the fourth quarter for each year is, despite 2001 and 2002 being bad years, there was substantial recovery. In 2003, the fourth quarter was a continuation of the strength of the previous three quarters (refer to Tables 1-3). However, the tables don’t show the day-to-day price action and the risk you may face by simply owning the top-five ETFs from the first two weeks. Figures 5, 6 and 7 show the day-to-day relative performance of the top-five ETFs and QQQ and SPY. The day-to-day percentage changes are tracked and adjusted to a start price of zero on the last day of the previous quarter. Figure 5 shows BHH was a steady climber in the fourth quarter of 2001. The remaining ETFs followed suit, but at a slower rate of ascent. In the fourth quarter of 2002 (see Figure 6), the ETFs peaked in early December and then declined. Figure 7, fourth quarter for 2003, was more volatile, with a sharp descent in the middle of November, and then a recovery. This study indicates if broad-based buying or selling in the top-volume

www.activetradermag.com • June 2004 • ACTIVE TRADER

ETFs occurs in the first two weeks of the FIGURE 7 4TH QUARTER 2003

quarter — accompanied by confirmation

Although follow-through from the beginning to the end of a quarter may be more likely when there is early-quarter strength among all ETFs, the price action can be wide ranging. The fourth quarter of 2003 was volatile, with a sharp drop in the middle of November, followed by a recovery.
40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% -5.00% HHH EWM EWY EWZ SMH SPY QQQ

from the QQQs and SPYs — that trend can persist to the end of the quarter. However, as Figures 5, 6 and 7 suggest, what happens along the way can be wide ranging.

Starting at the bottom
There are many ways to screen for topperforming ETFs. If the trend indicated in the first two weeks of a quarter is supported by broad-based participation, there is a better chance the top-performing ETFs in the early part of the quarter will follow-through in the same direction the remainder of the quarter. Traders looking for ETFs with the best odds for long-side trades can use a list of top performers as their departure point. If the Street has latched on to certain sectors, they are likely to benefit the most.Ý

Source: data - eSignal; chart - Excel

ACTIVE TRADER • June 2004 • www.activetradermag.com

33

FUTURES

&

OPTIONS

Trading Strategies

THE Market Facilitation INDEX
Traders are always looking for ways to determine when the market is poised to make a move. The Market Facilitation Index is a tool you can use to identify market lulls that can precede price trends.
BY THOM HARTLE
lancing at intraday charts shows the market often trends, either right from the opening or at some other time in the day, and all a trader has to do is be aggressive to take advantage of such moves. But more often than not, trying to capture intraday moves that have already begun is fruitless. As soon as you enter the trade, the market stalls and you are either pushed out by risk management rules or forced to wait, stressfully, for a resumption of the trend. If the market continues to languish, most traders will exit the trade — only to see the market begin to trend again. Nothing is more frustrating than having accurately anticipated the market’s direction and having nothing to show for it. What would help avoid these situations is an indicator that can signal lulls in the market. Quiet periods often precede trends, and by being able to identify potential trending periods

G

before they begin, a trader can get in the market quickly enough to capitalize on the maximum portion of a price move. One indicator for measuring whether the market is in a potential pre-trend lull is the market facilitation index (MFI). Here, we will analyze the price action after high MFI readings to see if there’s a relationship worth trading.

The indicator
Bill Williams, Ph.D., detailed the MFI in his book Trading Chaos (John Wiley & Sons, 2004). The MFI is a price bar’s range divided by its volume: MFI = (High – Low)/Volume The MFI will climb if volume subsides relative to the price bar’s range. Low volume activity is one way to identify a market that is pausing, a condition that can precede a trend. The raw MFI number is multiplied by 1 million because the base calculation is typically very small. For example, if a bar’s high for the E-mini S&P 500 futures (ES) was 1,133 and the low was 1,130, the range is 3 points. If the trading volume was 92,000, the raw MFI would be 0.00003261. Multiplying by 1 million results in 32.61, which is easier to work with.

FIGURE 1 THE MARKET FACILITATION INDEX (MFI) This 45-minute chart illustrates the tendency of the MFI to increase when the market trades sideways and falls into a lull. The MFI began to drop as the market moved into a downtrend.
September2004 E-mini S&P 500 futures (ESU04), 45-minute A B D E F C G 11,400.0 11,350.0 H I

The study
The following study used 45-minute bar data for the E-mini S&P 500 futures from Jan. 2, 2004, to June 10, 2004, a total of 111 days. Only the pit session hours were tracked, opening at 8:30 a.m. (CT) and closing at 3:15 p.m. For this period the mean (average) MFI reading was 63.16 and the standard deviation (STD) of all the readings was 23.98, meaning the MFI should fall between roughly 49 (the mean minus the STD) and 87 (the mean plus the STD) 68 percent of the time. Our goal is to see how the market behaves following a high MFI intraday value, which we will define as 87. This value (the mean plus one standard deviation) was chosen to generate more observations than would be the case if the mean plus two standard deviations were used. Also, 45-minute bars were chosen because as the time frame gets shorter the MFI becomes increasingly volatile and does not serve the purpose of identifying market lulls that should precede trends. Figure 1 is a 45-minute chart of the September E-mini S&P 500 futures contract (ESU04) with MFI and volume. We will walk through the chart bar by bar to better understand the MFI readings relative to the price action.
www.activetradermag.com • October 2004 • ACTIVE TRADER

Market facilitation index 100 50 a b Volume 6/28 Source: eSignal c d e f 0 6/29 g h i 100K

34

TABLE 1 OUT TO LUNCH The majority of high MFI readings during the analysis period occurred during the 11:30 a.m. and 12:15 p.m. bars. The market tended to trade sideways Time 8:30 9:15 10 10:45 11:30 12:15 1 1:45 for the first half of the day and as the vol(Central) a.m. a.m. a.m. a.m. a.m. p.m. p.m. p.m. ume decreased, the MFI rose. Bar A, the MFI > 87 0 0 6 27 47 54 20 4 first bar of the day, gapped up from the previous day’s close. The volume (histogram bar a) was nearly 100,000 conTABLE 2 FAVORABLE AND UNFAVORABLE MOVES tracts, producing an MFI of 32.70. The next bar (B), which AFTER HIGH-MFI BARS started at 9:15 a.m., was an inside bar (lower high and higher low than the preceding bar) and the volume dropped off to The median MFEs and MAEs following high-MFI bars just under 55,000 contracts (b). Consequently, the MFI rose to Occurrences Time MFE MAE 50.16. Bar C’s range was less than bar B’s and the volume trailed off to less than 38,000 contracts (c), and the MFI edged 6 10 a.m. 3.125 -0.875 up to 52.86. At bar D, the range expanded but the volume 27 10:45 a.m. 1.500 -0.500 declined to 24,000 (d) and the MFI climbed to 106.82. 47 11:30 a.m. 0.750 -2.000 Bar E, which began at 11:30 a.m., was another inside bar and 54 12:15 p.m. 1.750 -1.125 the volume dropped to slightly below 16,000 (e), pushing the MFI to 109.59. Bar F had a larger range but the volume only 20 1 p.m. 1.250 -1.500 reached 25,000 (f), so the MFI increased to 117.32. Note these 4 1:45 p.m. 1.375 -0.750 two bars (E and F) cover the time period most traders in Chicago and New York would be at lunch. FIGURE 2 FAVORABLE MOVES: COMPARING THE 11:30 AND 12:15 BARS Bar G was a wide-range down bar and the volume hit its peak for the day The maximum favorable excursions (MFEs) for the 11:30 a.m. bars (green) at nearly 145,000 contracts (g). The MFI are less than those for the 12:15 p.m. bars (blue) except for three occurdropped to 84.21. Bar H and bar I had rences. narrower ranges, but the volume remained high (near 80,000) and the MFI 8.00 dropped back to 37.98 and 40.71, respec7.00 tively. 11:30 bars 6.00 Figure 1 illustrates a number of attributes of the relationship between market 5.00 12:15 bars action and the MFI. The first bar gapped 4.00 higher, had high volume and a low MFI 3.00 reading. In this case, the high volume was less of a sign of good demand than 2.00 an indication demand was spent, as 1.00 there was no follow-through price 0 movement to the up side. (This is not always the case, as there are plenty of Number of occurrences days a market trends immediately from Source: Excel the opening bar, accompanied by heavy volume.) As the market traded sideways and the MFI rose to well above 87, the period. This suggests the MFI action in Figure 1 is typical. High market did, in fact, fall into a lull. Then, as it moved into a MFI readings, which imply low volume relative to the price downtrend, the MFI began to fall. The final two bars had high range, are followed by low MFI readings, which imply high volume, sideways price trend and high MFI readings. volume relative to the price range. Let’s look at the data for the entire study to see if we can The next question is what happens after high MFI readings? identify any consistencies in the price action and the MFI in We will look at the price action following MFI readings above terms of the time of each bar. 87 at the close of each 45-minute bar. To do this, we need a way to qualify which way the market moves in the bar immediateHow high, MFI? ly after each high MFI bar. Table 1 shows the number of times the MFI exceeded 87 for the Let’s say if a high MFI bar closes in the upper 50 percent of different 45-minute bars throughout the trading day. The MFI its range, the difference between the close of this bar and the never exceeded 87 during the first two bars, which encompass next bar’s high is the maximum favorable excursion (MFE) and the first 90 minutes of trading. Next, the number of high MFIs the difference between the close and the next bar’s low is the climbed until peaking at 54 during the 12:15 p.m. bar; the num- maximum adverse excursion (MAE). If a high MFI bar closes in ber of occurrences trailed off the rest of the day. the lower 50 percent of its range, the MFE is the difference The 54 occurrences equal just less than 50 percent of the total between the close and the next bar’s low; the MAE is the difnumber of times the MFI exceeded 87 over the 111-day study ference between the close and the next bar’s high. We are only
Maximum favorable excursion 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 ACTIVE TRADER • October 2004 • www.activetradermag.com 35

FUTURES

&

OPTIONS

Trading Strategies continued
Let’s take a closer look at the difference between the 11:30 a.m. bar and the 12:15 p.m. bar. Figure 2 shows the MFEs for the two periods sorted from smallest to highest. The MFE was zero for both bars four times. Notice how the MFEs for the 11:30 bars (green) trail below the MFEs for the 12:15 bars (blue) except for three occurrences. Figure 3 shows the MAEs for the two time bars sorted from smallest to largest negative move. The MAEs for the 11:30 a.m. bars are consistently worse than those for the 12:15 p.m. bars. (By contrast, the 12:15 p.m. bars with MFI readings below 87 had a median MFE of 1.50 and a median MAE of -2. The 64 times the MFI was below 87 during the 11:30 a.m. bars, the median MFE was 1.50 and the median MAE was 1.25.) Also, recall the price action in Figure 1 showed the 8:30 a.m. bar had no follow through that day after it closed and the MFI was below 40. How common is this? Figure 4 shows the MFEs and MAEs for the one hundred and eleven 8:30 bars with MFI readings below 87, sorted by MFE. Not until bar 66, which spans approximately 60 percent of the observations, is the MFE 1.75 or higher, which is the median MFE for the 12:15 p.m. bar. The median MFE for the bar following the 8:30 a.m. bars is 1.50 and the median MAE is -2.00.

FIGURE 3

UNFAVORABLE MOVES

Reflecting the tendency shown in Figure 2, the maximum adverse excursions (MAEs) of the 11:30 bars are consistently worse than those of the 12:15 bars. The 12:15 bars had larger favorable moves and smaller unfavorable moves than the 11:30 bars.
0 Maximum adverse excursion -2.00 -4.00 -6.00 -8.00 -10.00 -12.00 -14.00 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 -16.00 12:15 bars 11:30 bars

Number of occurrences Source: Excel

FIGURE 4

8:30 A.M. BARS

This chart shows the MFEs and MAEs for the one hundred and eleven 8:30 a.m. bars with MFI readings below 87, sorted by MFE.
10.00 8.00 8:30 a.m. bar MFE and MAE 6.00 MFE 4.00 2.00 0 -2.00 -4.00 -6.00 -8.00 1 6 11 16 21 26 31 -10.00 MAE 36 41 46 51 56 61 66 71 76 81 86

No free lunch
The study shows there is a tendency for the E-mini S&P 500 futures market to go into a low-volatility state, or lull, during lunchtime in Chicago. But after this lull the market is prone to trend. It appears taking a countertrend approach prior to lunch would be more profitable because of the lack of follow-through price movement during the lunch period. Also, the likelihood of a persistent trend early in the morning is not as likely following lunch. This research can be the basis for developing a strategy based on a shorter time frame, such as fiveminute bars. For example, if the MFI indicates a lull period during the 12:15 bar, use your strategy with more-distant targets than if you were placing a trade in the morning. Countertrend trading, or trading with very small targets, appears to be a better approach going into the lunch period. Trading is a probability issue. When we review charts with indicators and price patterns outlined our eyes are drawn to the great trades where the outcome is known after the fact. Without digging into the data it’s difficult to identify a particular market characteristic that can be the basis for a viable trading method. Once you find a characteristic, you can then develop a strategy around that information.
91 96 101 www.activetradermag.com • October 2004 • ACTIVE TRADER 106

Number of occurrences Source: Excel

checking the next bar’s behavior. Table 2 shows the median MFE and MAE for each occurrence of the MFI exceeding 87 by time and based on the combination of up- or downtrend movement. We used the median to reduce any skewing of the data by outliers. Although it has the highest MFE, the 10 a.m. bar only has six high MFI readings, so you cannot read too much into it. The 11:30 a.m. bar is more interesting. It starts at 11:30 a.m. and ends at 12:14:59 p.m. The median MFE is only 0.75 and the median MAE is -2, which suggests taking a directional trade going into lunchtime is a low-probability proposition. The 12:15 p.m. bar, which has the greatest number of occurrences, also has the highest MFE of all bars besides the 10 a.m. bar.
36

TRADING Strategies

The Squat Bar
Traders often anticipate trend moves to follow low-volume, low-volatility periods. The squat bar pattern shows high-volume bars can also lead to tradable price moves.
FIGURE 1 SQUAT BAR EXAMPLES A squat bar (highlighted red) has both higher volume and a lower MFI reading than the preceding bar.

BY THOM HARTLE

Nasdaq 100 index-tracking stock (QQQ), daily

35.00

34.00 raders have long studied the relationship between A price and volume looking for some definitive way to identify high-probability trades. For 33.00 example, traders sometimes look for low-volume trading ranges because of the theory that with such markets many participants are on the sidelines, and if some event triggers a price move, these MFI 0.800 sidelined players will react and potentially create a trend. 0.600 Many techniques and indicators, such 0.400 as on-balance volume and equivolume charts, have been devised specifically to Volume 1,500,000 gauge volume and price action. “The 1,250,000 1,000,000 Market Facilitation Index” (Active Trader, 750,000 October 2004 ) discussed another August volume-based indicator developed by 26 2 9 16 23 Bill Williams, who detailed the MFI in Source: CQGNet Inc. his book Trading Chaos (John Wiley & Sons, 2004). Williams used the MFI as one of two parameters to classify price bars into four cate- a range of $0.46 range. Its daily volume was 81,401,200. The gories. Here, we’ll analyze the category Williams referred to as MFI for that day was: the “squat bar” to see if it signals trade opportunities. MFI = (0.46/81,401,200)*1,000,000 = 0.00565

T

MFI revisited

First, let’s review the MFI, which is a price bar’s range divided by its volume: MFI = [(High – Low)/Volume]*1,000,000 Multiplying by 1 million makes the numbers easier to work with because the raw MFI value is typically quite small. For example, on Aug. 31, 2004, the Nasdaq 100 index-tracking stock (QQQ) had a high of $34.12 and a low of $33.66, for
37

The MFI measures the amount of price movement for each share traded. If volume increases and the price movement (range) remains the same, the MFI decreases. If the price movement stays the same and volume decreases, the MFI increases.

Squat bar definition
The squat bar has two criteria: The current bar’s MFI reading must be less than the previous bar’s, and the current bar’s volwww.activetradermag.com • January 2005 • ACTIVE TRADER

FIGURE 2 NARROW-RANGE SQUAT BARS This chart supports a general contention that many trends end (within one to three bars from a top or bottom) with squat bars, but many squat bars are part of continuation moves. Squat bars with trading ranges smaller than the previous day’s range are marked with dots below them. However, these bars failed to signal trading situations.
Nasdaq 100 index-tracking stock (QQQ), daily 39.00

38.00

37.00

36.00

35.00

34.00

33.00

2004 February March 2 12 20 26 2 9 17 23 1 8 Source: CQGNet Inc.

15 22

April May June 1 12 19 26 3 10 17 24 1 7 14 21

July August 1 12 19 26 2 9 16 23

Sept. 1

ume must be greater than the previous bar’s. Figure 1 shows a daily bar chart of the QQQ and highlights several squat bars (in red). Because the MFI measures range relative to volume, the lower the MFI, the more volume per unit of range. If volume increases from one bar to the next, we can conclude there was more activity on this bar than the previous bar. This is a squat bar. Williams used the term “squat” to convey the idea the market was squatting and preparing to make a big jump because the increased level of activity on such bars reflected a battle between the bulls and the bears; the battle is intense because the bar has high volume and relatively little price movement. Williams stated almost every trend ends with a squat bar as one of the three top or bottom bars, but he also said not every squat bar signals the end of a trend. Figure 2 is a daily chart of QQQ from January to September 2004 with squat bars highlighted in red. The figure indicates Williams’ statement is accurate: Squat bars do often occur within one to three bars of the end of trends, but many others are followed by continuation moves. The squat bars marked with dots are those with trading
ACTIVE TRADER • January 2005 • www.activetradermag.com

ranges smaller than their preceding bars. The theory is if the squat bar’s range was narrower than the prior day’s range, that day was a tight battle between the bulls and bears. However, the chart doesn’t indicate these bars signal particularly favorable trade opportunities. If the squat bar does represent a battle between buyers and sellers, a short-term trading opportunity might exist because the day after a squat bar, one side of the market will be right and the other will be wrong. Those who are wrong will be forced to react and exit their positions, which could create a trend move. For example, if you sold short one day and the next day the market rallied, you’d probably take a loss at a certain point. Similarly, if you went long one day and the market headed lower the next, you might be forced to sell. This idea is the opposite of the concept mentioned earlier of low volume indicating market participants are on the sidelines awaiting some event to trigger a price move.

The study
To test this idea, we analyzed daily QQQ data from Jan. 2, 2003,

38

FIGURE 3 BREAKING THE SQUAT BAR’S HIGH When price exceeded the high of a squat bar the next day, there was more upside follow-through than downside movement.
1.00 0.80 0.60 0.40 0.20 0.00 -0.20 -0.40 -0.60 -0.80 Number of occurrences Source: Excel High broken

FIGURE 4 BREAKING BELOW THE SQUAT BAR’S LOW A move below a squat bar’s low tended to be followed by more downside price action than upside movement.
1.50 Low broken 1.00 0.50 0.00 -0.50 -1.00 -1.50 Number of occurrences Source: Excel

outside days (bars with lower lows and higher highs than their preceding bars). Finally, the market opened the next day within the range of the previous squat bar 58 times (68 percent). Because the market opened within the range of the squat bar nearly 70 percent of the time and failed to exceed its high or low only 11.8 percent of the time, these statistics suggest the market has a tendency to trend the day after a squat bar and take out the previous day’s high or low. For example, July 26, 2004, in Figure 1 (bar A) is a squat bar. The next day, the market opened within the squat bar’s range, its high was $0.30 above the previous day’s high and its low was $0.32 above the prior day’s low. Let’s review the data to see the range of outcomes on the day immediately following a squat bar. Figure 3 sorts the differences between a squat bar’s high and the following bar’s high when that bar traded above the squat bar’s high (blue bars). The purple bars represent the differences between squat-bar lows and the following bars’ lows, shown as negative numbers (i.e., a purple bar with a value of -0.40 means that bar’s low was 40 cents above the squat bar’s low). An outside bar will have positive values for both the high and the low (notice the eight purple bars above the x axis). For example, number 30 represents the price action on July 27, the day following Figure 1’s first squat bar. Its high is $0.30 above the x axis and the low is $0.32 below the x axis. Figure 4 shows the differences, sorted from smallest to largest, for days when the following bar’s low exceeded the squat bar’s low. The positive bars represent the difference between the squat bar’s low and the next bar’s low (i.e., a purple bar with a value of -0.50 means that bar’s low was 50 cents below the squat bar’s low). The negative bars represent the differences between the squat bar’s high and the following bar’s high. Figures 3 and 4 both show good price movement following a squat bar. If the high is broken, the average move is $0.26 above the squat bar’s high. If the low is broken, the average move is $0.39 below the squat bar’s low.

to Aug. 31, 2004. Of the 418 days tracked, 85 days (20 percent) were squat bars, which means they appeared roughly once a week. Squat bars were followed by inside days (bars with lower highs and higher lows than their preceding bars) only 10 times (11.8 percent of the time); eight squat bars were followed by

Using the squat bar’s probabilities
The bottom line of this study is the market remains within the squat bar’s range only 11.8 percent of the time. If the market opens within the squat bar’s range, which happens 68 percent of the time, there is a good chance the prior high or low will be broken. Therefore, you can set trade targets outside the squat bar’s range. If you get the direction right for your original position, the odds are good you’ll be able to capture the price move. The squat bar offers a unique way to look at price and volume action. Many trading setups are based on momentum and volume reaching low points. However, the squat bar represents the opposite approach, because its setup is based on volume hitting a high level. Ý

Additional Active Trader reading
“The Market Facilitation Index,” by Thom Hartle, October 2004, p. 70. You can purchase past articles at www.activetradermag.com/purchase_articles.htm and download them to your computer.
39

www.activetradermag.com • January 2005 • ACTIVE TRADER

TRADING Basics

HOLDRS: Stock groups in a single trade
Exchange traded funds (ETFs) such as HOLDRS allow traders and investors to specialize in certain indices, market sectors, or groups with a single “stock.”
Company Depositary ReceiptS. HOL- DRS with their respective 2004 percentDRS offer the opportunity to trade a age returns. For comparison, three of the group of stocks with a single instrument, most popular index-based ETFs — the and thereby profit if a particular indus- Nasdaq 100 tracking stock (QQQQ), the urrently, there are more try, group, or sector does well. The S&P 500 Depositary Receipts (SPY), and than 160 exchange-traded advantage of the group ownership rep- the Dow Industrial Diamonds (DIA) — funds (ETFs) traders and resented by a HOLDRS is you don’t also are included. The Market 2000+ HOLDRS (MKH) investors can analyze or have to pick the best stock. Few investtrade. These instruments, which trade ment decisions are as disappointing as represents 50 companies trading on the like common stock, represent a wide owning an individual company that is New York Stock Exchange, American range of market indices, sectors, and having problems while its industry is Stock Exchange, or the Nasdaq National groups. Here, we will look at one group doing well. (Of course, there’s always Market System that were considered to be of ETFs — Merrill Lynch HOLDRS — the risk that a single company’s woes among the largest in terms of worldwide market capitalization on July 7, 2000. The will negatively impact an entire group.) and review their salient features. Merrill Lynch has created 17 HOL- Europe 2001 HOLDRS (EKH) represents HOLDRS stands for HOLding DRS. Overall, there is a 50 of the largest European companies slight emphasis on the (worldwide market capitalization on TABLE 1 MERRILL LYNCH HOLDRS Internet, but groups such as Nov. 14, 2000) whose equity securities The 2004 percentage return of the HOLDRS banking and utilities are were listed for trading on a U.S. stock ranged from 42.08 percent (Internet) to –19.57 also represented. Table 1 market. percent (Semiconductor). Seven HOLDRS out perFigure 1 shows the HOLDRS’ 2004 lists the seventeen HOLformed the broader market indexes, such as the Nasdaq 100, the S&P 500, and the Dow Jones. FIGURE 1 2004 HOLDRS RETURNS

BY THOM HARTLE

C
Symbol HHH OIH UTH BDH WMH IIH BBH QQQQ RTH SPY RKH TTH SWH IAH EKH MKH BHH DIA PPH SMH

Name 2004 Return Internet HOLDRS 42.08% Oil Service HOLDRS 37.21% Utilities HOLDRS 24.59% Broadband HOLDRS 24.16% Wireless HOLDRS 22.54% Internet Infrastructure 21.69% Biotech HOLDRS 13.01% Nasdaq 100 9.46% Retail HOLDRS 9.18% S&P 500 8.62% Regional Bank HOLDRS 7.04% Telecom HOLDRS 6.19% Software HOLDRS 5.81% Internet Architecture 5.25% Europe 2001 HOLDRS 4.34% Market 2000+ HOLDRS 3.27% B2B Internet HOLDRS 2.89% DJI 30 2.81% Pharmaceutical HOLDRS -8.57% Semiconductor HOLDRS -19.57%

In this histogram bar format returns are sorted from lowest to highest. Seven of the HOLDRS outperformed the Nasdaq 100 tracking ETF (QQQQ). Utilities HOLDRS benefited from a low interest rate environment and Oil Services did well due to higher oil prices.
2004 percentage returns 50.0 40.0 2004 percentage returns 30.0 20.0 10.0 0.0 -10.0 -20.0 QQQQ TTH PPH EKH SPY RKH RTH SWH BBH UTH DIA BHH MKH SMH WMH BDH OIH IAH HHH IIH -30.0

Exchange traded funds Source: eSignal

Source: eSignal

40

www.activetradermag.com • April 2005 • ACTIVE TRADER

percentage returns sorted from lowest to highest. The Semiconductor HOLDRS (SMH) fared worst, posting a -19.57-percent loss for the year; the Internet HOLDRS (HHH) gained 42-percent. Seven HOLDRS outperformed the QQQQ, eight outperformed the SPY, and all but two outperformed the DIA. Let’s take a closer look at the two extremes — the Semiconductor HOLDRS and the Internet HOLDRS. Table 2 lists the top five companies (by current percentage weighting) held in the Semiconductor HOLDRS, as well as their respective 2004 returns. The Semiconductor HOLDRS represents companies that develop, manufacture, and market integrated circuitry and other products made from semiconductors for computers and other electronics. The largest company by weight, Intel (INTC), suffered the largest percentage loss in 2004 (-27.02 percent vs. -19.57 percent for the SMH itself). The diversification offered by owning the SMH provided a better return compared to an entire investment in Intel. Table 3 shows the opposite side of the coin. It lists the 2004 returns for the top five stocks (based on current weightings) in the Internet HOLDRS. Three of the top five had substantially higher returns than the Internet HOLDRS. (Amazon.com was actually down on the year.) Tables 2 and 3 demonstrate the double-edged sword of diversification. For traders, high liquidity is critical. Table 4 lists the average daily volume for the first five trading days of 2005. Three HOLDRS — Semiconductor (SMH), Oil Service (OIH), and Retail (RTH) — had average daily volume above one million shares for the first week of the year. At the other extreme is the Europe+ 2001 HOLDRS, which averaged less than 2,000 shares a day. Figure 2 is a ratio spread chart of the Internet HOLDRS (HHH) and the S&P 500 Depositary Receipt (SPY) — i.e., the closing price of HHH divided by the closing price of SPY. The chart highlights HHH's relative strength for most of 2004, as well as the sharp down moves that occurred in summer 2004 and at the beginning of this year. Figure 3 is a similar chart of the Oil Service HOLDRS (OIH) and SPY.

TABLE 2 TOP-FIVE SEMICONDUCTOR HOLDRS STOCKS Here is a list of the top five companies by percentage weighting held in the Semiconductor HOLDRS. All five showed negative returns for the year 2004. The Semiconductor HOLDRS was only down 19.57 percent while the largest company, Intel Corp., was down 27.02 percent. Stock Symbol Intel Corporation INTC Texas Instruments TXN Applied Materials AMAT Analog Devices ADI Maxim Integrated Products MXIM
Source: eSignal

Shares 30 22 26 6 5

Weighting 2004 Return 21.83% -27.02% 15.82% -16.20% 13.37% -23.80% 6.76% -19.12% 6.19% -14.50%

TABLE 3 TOP-FIVE INTERNET HOLDRS STOCKS Here is a list of the top five companies by percentage weighting held in the Internet HOLDRS. Four of the five showed positive returns with eBay Inc. up 80.07 percent, Yahoo! Inc. up 67.39 percent, and McAfee up 92.35 percent for the year 2004. The Internet HOLDRS was up just 42.08 percent for the year. Stock eBay Yahoo Time Warner Amazon.com McAfee
Source: eSignal

Ticker EBAY YHOO TWX AMZN MFE

Shares 24 52 42 18 7

Weighting 38.19% 28.03% 11.82% 11.76% 2.63%

2004 Return 80.07% 67.39% 8.12% -15.83% 92.35%

FIGURE 2 RELATIVE PERFORMANCE (HHH/SPY) This spread chart shows the price of the Internet HOLDRS divided by the price of the S&P 500 Depository Receipts (SPY). The Internet HOLDRS dominated SPY receipts, except during the market correction that spanned July through August and also during January, when eBay released negative news.
Internet HOLDRS/S&P 500 spread ratio (HHH/SPY), weekly 0.58 0.56 0.54 0.52 0.50 0.48 0.46 0.44 Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. 2005 Feb.

Source: TradeStation

ACTIVE TRADER • April 2005 • www.activetradermag.com

41

TABLE 4 AVERAGE DAILY VOLUME Here, the 5-day average volume for the first week of January 2005 is listed. The Semiconductors and the Oil Service HOLDRS demonstrate the highest daily liquidity. Symbol BBH BDH BHH EKH HHH IAH IIH MKH OIH PPH RKH RTH SMH SWH TTH UTH WMH
Source: eSignal

FIGURE 3 RELATIVE PERFORMANCE (OIH/SPY) The price of the Oil Services HOLDRS (OIH) is divided by the price of the S&P 500 Depository Receipts (SPY). Investors wanting to profit from the current rising price of oil can use the OIH. The price action of the OIH has steadily outperformed the SPY over the past year.
Oil services/S&P 500 spread ratio (OIH/SPY), weekly

5-day avg. volume 912,120 84,160 256,600 1,760 544,380 31,460 267,420 10,480 3,076,700 728,860 340,360 1,679,360 35,565,880 572,980 505,880 493,200 45,160

0.76 0.74 0.72 0.70 0.68 0.66 0.64 0.62 0.60 0.58 2004 Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. 0.56 Dec. 2005 Feb.

Source: TradeStation

OIH was also relatively strong in 2004 — buoyed by record oil prices — and unlike the HHH, its strength carried over into early 2005. These types of charts are good tools for understanding the behavior of a particular HOLDRS vs. the broader market.

Finding a theme
Unlike other ETFs, HOLDRS can only be bought and sold in round lots of 100 shares. As a result, they are not always an affordable option. For example, as of Feb. 4, the Biotech HOLDRS was trading at 145. 45, meaning a minimum investment in the BBH would be $14,545. By compar-

ison, other ETFs such as the QQQQ or the SPY can be bought and sold in any amount (e.g., 10 shares of the QQQQ, 35 shares of the SPY). Exchange traded funds are excellent tools for putting capital to work in specific equity groups, sectors, industries, as well as countries and fixed-income products. Tracking the performance of the HOLDRS can provide insight into what investment “themes” are working or not working, or which areas of the market are in or out of favor. Considering the rise in oil prices, it is not surprising the Oil Service HOLDRS was one of the top

performers in 2004, just as the Pharmaceutical HOLDRS (PPH) was one of the worst performers because the problems of Merck & Co. and Pfizer, Inc. For detailed information on HOLDRS performance, see the ETF Snapshot on p. 49. To see which HOLDRS are performing the best on weekly basis, refer to the Weekly HOLDRS review on the Active Trader Web site ( w w w. a c t i v e t r a d e r m a g . c o m ) . Additional information on HOLDRS can be found at www.holdrs.com.

ACTIVE TRADER • April 2005 • www.activetradermag.com

42

TRADING Strategies

DISSECTING T-NOTE FUTURES:

Tendencies and characteristics
BY THOM HARTLE

S

even years after the Chicago Board of Trade (CBOT) introduced 30-year U.S. T-bond futures in August 1977, the exchange launched the 10-year T-note contract (TY), which is now the most heavily traded treasury futures product in the U.S.

Ten-year T-notes are a sensitive measure of the fundamental forces driving interest rates. Any hints of inflation and the so-called “bond vigilantes” will sell bonds and notes, driving up interest rates. Conversely, if they see the economy softening, they will bid up prices for bonds and notes, pushing interest rates down.

Analyzing the price history of the 10year T-note futures market will allow us to identify its characteristics and determine if there has been any shift in its typical price behavior. We will identify typical daily ranges, closing prices relative to daily range, as well as tendencies for the low relative to the opening price

FIGURE 1 CONTINUOUS 10-YEAR T-NOTE FUTURES The 10-year T-note futures contract is the highest-volume U.S. treasury futures contract. Over the past year, the contract has been in uptrending, downtrending, and trading-range environments.
10-year T-notes (ZN), daily 112-00 111-00 110-00 109-00 108-00 107-00 106-00 105-00 104-00

Total

1,500,000 1,000,000 500,000

March 2004 April Source: CQGNet, Inc.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

2005

Feb.

43

www.activetradermag.com • July 2005 • ACTIVE TRADER

A detailed understanding of a market’s price history and characteristics allows you to craft trade strategies founded on statistical reality rather than casual observation. The following analysis takes the pulse of the T-note futures market.

for days the market closes up, and the high relative to the opening when the market closes down. With this information in hand, we’ll have a better understanding of this market and a stronger foundation on which to build strategies.

FIGURE 2 10-YEAR T-NOTE YIELD This chart of T-note yield covers the same time period as Figure 1 and shows the inverse relationship between price and yield.
10-year T-note yield 5.00 4.75 4.50 Yield 4.25 4.00 3.75 3.50 3/1/04 3/15/04 3/29/04 4/12/04 4/26/04 5/10/04 5/24/04 6/7/04 6/21/04 7/5/04 7/19/04 8/2/04 8/16/04 8/30/04 9/13/04 9/27/04 10/11/04 10/25/04 11/8/04 11/22/04 12/6/04 12/20/04 1/3/05 1/17/05 1/31/05 2/14/05 2/28/05 Source: Excel, Data: CQGNet, Inc.

A look at the charts
The review period spans March 1, 2004 to Feb. 28, 2005. Figure 1 is a continuous daily chart of the electronically traded T-note futures, which is the volume leading in the Treasury futures contract — daily volume commonly exceeds one million contracts. T-note prices and yields are inversely related. Figure 2 shows the T-note yield for the same period shown in Figure 1. Figure 1 shows three market phases have occurred over the past year: a downtrend, during which price dropped from above 112-00 to below 104-00, followed by an uptrend that took the market back above 112-00. Then the market settled into a trading range bounded roughly by 109-00 on the downside and 112-00 on the upside.

FIGURE 3 DAILY RANGES Daily ranges of the 10-year T-note contract have contracted over the past year, as highlighted by the down-sloping linear regression line.
Price has been converted to decimal format.

2.0

Daily ranges
Figure 3 plots the daily ranges of T-note futures. Keep in mind this market trades on the E-CBOT, the CBOT’s electronic arm, and is available globally 21 hours a day (opening each evening at 7:00 p.m. CT and closing at 4:00 p.m. the next day). There are two interesting characteristics. First, there is a monthly spike in the size of the daily ranges and, second, the size of the daily ranges has been trending down over the past year.
1.5 Points 1.0 0.5 0.0 1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 44

Source: Excel, Data: CQGNet, Inc.

ACTIVE TRADER • July 2005 • www.activetradermag.com

FIGURE 4 MOST FREQUENTLY RECURRING DAILY RANGES This distribution of daily ranges shows how much you can expect T-notes to move on a given day based on the review period. The daily range was less than 24/32nds on 205 days, or 82 percent of the time.
Frequency distribution of daily ranges 70
Price has been converted to decimal format.

Bond and note basics
reasury bonds and notes are debt securities issued by the U.S. Treasury. Bonds and notes are considered debt because by purchasing these instruments you are, in effect, loaning money to the Treasury department, which then pays you interest (determined by the “coupon rate”) on a semiannual basis and returns the principle when the bond or note matures on the maturity date. For example, if you purchased a $1,000 10-year T-note with a 4-percent coupon, you would receive $20 every six months, totaling $40 per year and the $1,000 would be paid back to you on the maturity date 10 years from now. T-bonds and T-notes are called “fixed-income” securities because of the fixed coupon payment an investor receives while holding the bond or note. T-notes are issued in maturities of two, three, five, and 10 years; T-bonds have maturities greater than 10 years. (In October 2001, the Treasury department suspended issuing fixed-principle Treasury bonds, but there is still a large amount of unredeemed bonds traded in the open market.) T-bonds are priced as a percentage of the notational value with the combination of the handle, 100 for example, and 32nds of 100. For example, 98-14 is a price that trans-

60 50 Frequency 40 30 20 10 0 1.50 1.75 2.00 2.25 2.50 2.75 0.00 0.25 0.375 0.50 0.625 0.75 0.875 1.00 1.125 1.25 3.00

T

Source: Excel, Data: CQGNet, Inc.

FIGURE 5 WHERE DOES THE T-NOTE CONTRACT TEND TO CLOSE? This depiction of where the T-note closed relative to the day’s range indicates T-notes tend to close more toward the daily high or low — i.e., below the 30-percent level of the range or above 70-percent level of the range.
45 40 35 30 Frequency 25 20 15 10 5 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% 80.00% 90.00% 100.00% 0

Source: Excel, Data: CQGNet, Inc.

Plotting a linear regression line (see “Key Concepts and Definitions” on p. 92) through the data highlights the downward trend. The formula for the linear regression trendline shows its first value is nearly 24/32nds and its last value is below 16/32nds. The monthly one-day spike in the ranges is the result of the release of employment data the first Friday of each month. Although this event currently has
45

more impact on prices than any other economic release, this may not always be the case. There was a time when money supply numbers dominated treasury crowd psychology. In the future, it could be price or inflation measurements that influence traders’ actions. The decline in daily ranges could be the result of the trading range the market moved into during the last three months of the analysis period (trading

ranges are often accompanied by decreased volatility). If you trade with preset targets, you need to downsize these target price levels when daily ranges are contracting. If the daily ranges begin to expand again, you can set targets farther from your entry level. Figure 4 is a frequency distribution of the daily ranges, which provides an indication of how much you can expect T-notes to move on a given day. The market had a range larger than 16/32nds (.50) but less than 20/32nds on

www.activetradermag.com • July 2005 • ACTIVE TRADER

FIGURE 6 OPEN-LOW DIFFERENCE Of the 124 times the market closed up for the day, it also closed below the open only eight times, and only twice at the opening price. lates to 98-14/32nds or $984.38 for a $1,000 T-bond. T-notes are priced in a similar fashion, except they can trade in one half of a 32nd, which is also referred to as “plus.” For example, 98-14+ is 98-14.5/32nds. As market conditions change, so do fixed income players’ expectations. For example, if there are signs of inflation, the purchasing power of the fixed cash flow from coupon payments would diminish. Consequently, bond and note investors would demand a higher yield to be compensated for the lower future purchasing power. In the next Treasury auction, investors might be interested in buying 10year T-notes only if they receive a 5 percent coupon. Also, if you wanted to sell a 4-percent coupon T-note, you would have to mark the price down from 100 to 80 (mirroring the 20-percent difference in its coupon rates and a 5-percent coupon Tnote). This is why bond and note prices fall when yields rise. For many investors, Treasury bonds and notes are appealing because they are backed by the full faith and credit of the U.S. Treasury, which means your principle and coupon payment have virtually no risk. This is very different from owning a stock, for which there is no guarantee the company will be profitable in the future.
2.5 2.0 1.5 1.0 Points 0.5 0.0 -0.5 -1.0 -1.5 Source: Excel, Data: CQGNet, Inc. Low for days T-note closed up
Price has been converted to decimal format.

FIGURE 7 OPEN-LOW FREQUENCY DISTRIBUTION Here’s the frequency distribution of the differences from Figure 6. Of the 124 times the market closed up, the low was less than 8/32nds below the open 105 times. (The market traded more than 16/32nds below the open and still closed up for the day only once.) If you’re long and the market is more than 8/32nds below the opening and below the previous day’s close, the odds of a recovery are low.
Open to low for up closes 35
Price has been converted to decimal format.

30 25 Frequency 20 15 10

64 days (the peak reading), or 25 percent of the time. The daily range was less than 24/32nds on 205 days, or 82 percent of the time.

5 0 0.000 0.125 0.250 0.375 0.500 0.625 0.750 0.875 1.000 1.125 1.250 1.375 1.500 46

Source: Excel, Data: CQGNet, Inc.

What does the closing price indicate?
Next, Figure 5 highlights tendencies of the daily close relative to the day’s range. The chart is a frequency distribution of where the T-note futures closed relative to the day’s range on a percentage basis. The e-CBOT T-note for settle-

ment purposes uses the pit closing price, which closes at 2:00 p.m. CT and the electronic session trades for two more hours. For example, the market closed between the 20- and 30-percent levels of the day’s range 40 times (third bar from the left).

Overall, the chart indicates T-notes tend to close more toward the daily high or low — i.e., below the 30-percent level of the range or above the 70-percent level of the range. The largest number of closes (41 occurrences) fell in the 80- to

ACTIVE TRADER • July 2005 • www.activetradermag.com

FIGURE 8 OPEN-HIGH DIFFERENCE The raw differences between the open and the high on days the market closed down has a subtle difference from Figure 6: The high is more than 16/32nds above the open numerous times.
High for days T-note closed down 1.5 1.0 0.5 0.0 Points -0.5 -1.0 -1.5 -2.0 -2.5
Price has been converted to decimal format.

-3.0 Source: Excel, Data: CQGNet, Inc.

90-percent level (second bar from the right). The market closed in the 50 to 60percent level of the daily range only 23 times, or less than 10 percent of all days. The lesson from this data is that traders should consider using rules that enable them to stay with the trend for the day.

Another attribute to consider is how the market trades on days it closes up vs. days it closes down. Figure 6 displays the difference between the opening price and the low for days when the market closes higher than the previous day’s close. Of the 124 times the market

FIGURE 9 OPEN-HIGH FREQUENCY DISTRIBUTION The frequency distribution of the data from Figure 8 shows 96 of the 121 days the market closed down, the high was less than 8/32nds above the open. However, on seven occasions the market traded more than 16/32nds above the open and still closed down for the day. This tendency suggests long traders must be wary even when the T-note is up more than 16/32nds above the opening price.
35 30 25 High for days T-note closed down
Price has been converted to decimal format.

20 15 10 5 0 0.000 0.125 0.250 0.375 0.500 0.625 0.750 0.875 1.000 1.125 1.250 1.375 1.500

closed up for the day, the market closed below the open only eight times, and just twice the market closed at the opening price. Figure 7 is the frequency distribution of the differences from Figure 6. Of the 124 times the market closed up, the low was less than 8/32nds below the open 105 times (85 percent). Only one time did the market trade more than 16/32nds below the open and still close up for the day. What does this tell us? If you are long and the market is more than 8/32nds below the opening and below the previous day’s close, your chances of a recovery are not very good. Reversing the perspective and looking at the difference between the open and the high for those days the market closed down leads to Figures 8 and 9. Figure 8 plots the raw differences between the open and the high. There is a subtle shift in the data: The high is more than 16/32nds above the open numerous times. Figure 9 is a frequency distribution of the data from Figure 8. It shows that in 96 of the 121 days (79 percent) the market closed down, the high was less 8/32nds above the open. However, on seven occasions the market traded more than 16/32nds above the open and still closed down for the day. This tendency suggests traders tend to be optimistic about a market holding onto gains, and long traders must be wary even when the T-note is up more than 16/32nds above the opening price. The market traded above the opening price and closed down for the day only three times. The T-note closed unchanged only five times during the review period. One big-picture observation: Figure 10 is the long-term view of the 10year T-note yield (logarithmic scale). Following the rate peak in 1982, yields have steadily declined for more than 20 years — the kind of long-term trend often referred to as a “secular” move. Once this long-term trendline is broken, fixed income traders will be watching for any signs the interest rate market is

Source: Excel, Data: CQGNet, Inc.

Frequency

47

www.activetradermag.com • July 2005 • ACTIVE TRADER

Related reading
Related Active Trader articles by Thom Hartle:
“Treasury bonds and notes,” Active Trader, June 2005. A primer on the bond and note market, covering the cash market, bond funds, and ETFs and futures. “Short-term T-bond trading,” Active Trader, October 2002. A strategy that uses a multiple time frame approach: Two indicators applied to daily bars work together to determine the trend; two others, Bollinger Bands and the moving average convergence-divergence (MACD) indicator, identify entry and exit signals on an intraday basis. “Relative volume analysis,” Active Trader, July 2003. Traders commonly use total volume to take the pulse of price moves, but when it comes to determining whether the bulls or bears are in charge of the market, up and down volume speaks volumes. “Familiarity breeds profitability,” Active Trader, September 2002. The author analyzes price patterns to determine the odds that different kinds of price moves will occur. “Up-down volume and next-day follow-through,” Active Trader, December 2004. What kind of trading is likely to occur today? Yesterday’s balance of up volume and down volume — and whether the market establishes the high or low of the day first — provides guidance in the Nasdaq 100. “Getting in on follow-through days,” Active Trader, January 2004. In a follow-up to the previous article’s discussion of the odds of next-day followthrough in the S&P futures, the author looks at the realities of basing entries on this price behavior. “Following through in the S&Ps,” Active Trader, December 2003. Strong closes and large ranges are often interpreted as signs of potential followthrough. This study unveils another way to find out what today’s market action says about tomorrow by analyzing the NYSE updown volume statistics at the close of the day to see if there is any consistent followthrough price action in the E-mini S&P 500 futures the next day. You can purchase and download past Active Trader articles at www.activetradermag.com/purchase_articles.htm.
48

FIGURE 10 T-NOTE BIG PICTURE Yields have steadily declined since the rate peak in 1982. When this longterm trend is broken, fixed income traders will be watching for any signs the interest rate market is moving again into another secular rising-rate trend.
20 15 10 CBOE 10-year yield as of 3/24/05

Yield

5 3

1965

1970

1975

1980

1985

1990

1995

2000

Source: Excel, Data: CQGNet, Inc.

FIGURE 11 EMPLOYMENT FRIDAY: ONE-MINUTE CHART With expectations for a rise of 200,000 non-farm payroll jobs, the number came in at 146,000. The T-note contract started falling before the number’s release at 7:30 (CT), but rallied when the news actually came out.
10-year T-note (ZN), One-minute 110-00

109-28

109-24

109-20

109-16

109-12 Volume 15,000 10,000 5,000 0 7:10 7:15 7:20 7:25 7:30 7:35 7:40 7:45 7:50 Source: CQGNet, Inc.

moving again into a secular rising-rate trend, such as the one that ended in 1982.

T-notes and the employment report
The Commerce Department releases the employment report the first Friday of each month. The treasury market’s initial

reaction to the report is typically very volatile. Because employment-report Fridays are such exceptional trading days, it makes sense to analyze them separately from the rest of the T-note data. Figure 11 is a one-minute T-note chart following the release of the Feb. 4, 2005 employment report. Expectations were

www.activetradermag.com • July 2005 • ACTIVE TRADER

FIGURE 12 RANGES FOR EMPLOYMENT FRIDAYS Similar to the overall T-note data in Figure 3, the size of the Employment Friday daily ranges has been declining.
3.0 2.5 2.0 Points 1.5 1.0 0.5 3/1/04 4/1/04 5/1/04 6/1/04 7/1/04 8/1/04 9/1/04 1/1/05 10/1/04 11/1/04 12/1/04 2/1/05 0 Employment
Price has been converted to decimal format.

Source: Excel, Data: CQGNet, Inc.

FIGURE 13 T-NOTE RANGES WITHOUT EMPLOYMENT FRIDAYS Removing the Employment Fridays from the data decreases the typical daily ranges. The linear regression line is shifted slightly lower, but the slope is nearly the same.
2.0
Price has been converted to decimal format.

1.5 Points 1.0 0.5 0.0 1 12 23 34 45 56 67 78 89 100 111 122 133 144 155 166 177 188 199 210 221 232

Source: Excel, Data: CQGNet, Inc.

for a rise of 200,000 non-farm jobs, but the number came in at 146,000. Price actually started falling before the number came out at 7:30 (CT), and then rallied when the news was released. As the price peaked above 110-00, three one-minute bars had volume in excess of 15,000 contracts — that’s 250 contracts ($25 million) changing hands per second. Figure 12 displays the ranges for the 12 employment Fridays in the analysis period (Figure 11 represents the last Friday in Figure 12). Similar to the overall data shown in Figure 3, the size of the employment Friday ranges has been declining. If we remove the employment Fridays from the data, the typical daily range drops. Figure 13 is the same as Figure 3 except that employment Fridays are removed. The linear regression line is shifted slightly lower, but the slope is nearly the same (-.0009 versus -.001). Figure 14 is the frequency distribution of daily ranges without the employment days. Only one day exceeded a range of 1.5 points, vs. 11 additional when employment Fridays are included (see Figure 4). Obviously, it’s a good idea for T-note traders to develop two sets of trading procedures — one for employment Fridays, and the second for the rest of the month.

FIGURE 14 RANGE DISTRIBUTION FOR EMPLOYMENT FRIDAYS Without Employment Fridays, only one day in the review period had a range larger than 1.5 points. This highlights the importance of developing two sets of trading procedures in T-notes — one for employment Fridays, and the second for the rest of the month.
70
Price has been converted to decimal format.

Taking a market’s pulse
Performing relatively simple statistical analysis can give you a clearer understanding of a market’s typical behavior and provide the foundation for new trading approaches, or improvements to existing ones. Comparing recent price action to that of a year ago will alert you to, among other things, volatility changes that critically impact trading decisions and profit potential. Also, separate analysis of a market’s reaction to specific economic reports can alert you to the best ways to trade in these unique situations. As the discussion regarding employment data indicates, the T-note market is very different on employment Fridays than other trading days.

60 50 Frequency 40 30 20 10 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 0

Source: Excel, Data: CQGNet, Inc.

49

www.activetradermag.com • July 2005 • ACTIVE TRADER

TRADING Strategies

The QQQQ/SPY spread
BY THOM HARTLE

50

www.activetradermag.com • September 2005 • ACTIVE TRADER

Analyzing the relative strength of major stock indices can help you define the market’s current condition and exploit trade setups at potential turning points.

D

uring bull markets money flows into growth-oriented stocks as investors attempt to capitalize on companies with the potential to expand in the favorable economic environment.

When the economic outlook sours, investors will move money out of growth

stocks and into so-called “defensive stocks” — those companies that offer products and services consumers tend to purchase despite the weak economy, and which often pay dividends. These stocks do not necessarily go up in a weak environment, but they tend to fall less than the growth-oriented companies. Comparing growth and defensive stocks is one way to characterize the broader market environment and trend: When growth stocks are leading defensive stocks, the outlook for the broad market is good. If defensive stocks are outperforming, the outlook is negative. A relative strength calculation of two of the most popular exchange traded funds (ETFs) — the Nasdaq 100 index tracking stock (QQQQ) and the S&P 500 Depositary Receipts (SPY) — allows you to gauge the stock market’s overall trend and set up swing trades that are in sync with that trend. The setup outlined here uses an entry concept based on the work of Richard Wyckoff, a stock market analyst and educator from the 30s.

ACTIVE TRADER • September 2005 • www.activetradermag.com

51

FIGURE 1 QQQQ/SPY RATIO Dividing the Nasdaq 100 tracking stock (QQQQ) by the S&P 500 Depositary Receipts (SPY) reflects how the QQQQ is performing relative to SPY. A rising QQQQ/SPY ratio indicates money is flowing into technology stocks; a weak QQQQ/SPY ratio suggests funds are flowing out of technology and into more “defensive” S&P stocks.

highlights how this ratio has tended to trend: In the fourth quarter of 2004, the ratio was in an uptrend, while during Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts the first part of 2005 it was trending (QQQQ/SPY), daily down. What does the trend of the QQQQ/SPY ratio tell us? To help 0.33 answer that question, let’s break down the components of these two ETFs. Table 1 shows the top 10 holdings and industry groups in the S&P 500. The S&P 500 0.32 contains a high percentage of defensive stocks, which are those companies that tend to remain stable when the economy and the stock market are weak. 0.31 Industries that provide products and services consumers need even when the economy is moving into a recession include financials, health care, consumer staples, and utilities. These four groups Aug. Sept. Oct. Nov. Dec. 2005 Feb. March April May make up 45 percent of the S&P 500. Table 2 lists the top 10 holdings and Source: CQGNet industry groups for the Nasdaq 100. In is to calculate the ratio between two contrast to the S&P 500, the Nasdaq 100 Comparing indices contains no financial stocks, and health Relative strength is a comparison of one markets. Figure 1 shows the price of QQQQ and consumer goods make up less than market’s performance to another’s. A simple way to measure relative strength divided by the price of SPY. The chart 4 percent of the index. Information technology hardware and computer software/services comprise TABLE 1 TOP 10 S&P 500 HOLDINGS AND INDUSTRY GROUPS* more than 55 percent of the index. Top 10 holdings % of Top 10 industry groups % of Given its makeup, the Nasdaq index index 100 should lead the S&P 500 1 General Electric Co. 3.41 1 Financials 20.73 when the economic outlook is 2 Exxon Mobil Corp. 3.02 2 Information technology 14.97 favorable and funds flow into growth-oriented technology 3 Microsoft Corp. 2.89 3 Consumer discretionary 13.33 companies. If the economic/ 4 Pfizer Inc. 2.22 4 Health care 13.03 stock market backdrop is nega5 Citigroup Inc. 2.20 5 Industrials 12.64 tive, the Nasdaq 100 index 6 Wal-Mart Stores Inc. 2.17 6 Consumer staples 8.22 should lead the S&P 500 down as 7 American Intl. Group Inc. 1.70 7 Energy 7.19 funds flow out of growth com8 Bank of America Corp. 1.69 8 Telecommunication services 3.69 panies and into the S&P 500 9 Johnson & Johnson 1.61 9 Materials 3.15 defensive stocks. 10 International Business Machines 1.38 10 Utilities 3.05 The market adage that “a rising tide lifts all ships” is central Total 22.29 Total 100.00 to this concept. If the Nasdaq 100 *As of 09/30/2004 is headed up and the S&P is lagSource: Nasdaq.com

52

www.activetradermag.com • September 2005 • ACTIVE TRADER

TABLE 2 THE TOP 10 HOLDINGS AND THE INDUSTRY GROUP BREAKDOWN OF THE NASDAQ 100 INDEX* Top 10 holdings 1 Microsoft Corp. 2 Qualcomm Inc. 3 Intel Corp. 4 Apple Computer Inc. 5 Cisco Systems Inc. 6 Nextel Communications Inc. 7 eBay Inc. 8 Dell Inc. 9 Amgen Inc. 10 Comcast Corp. Total *As of 03/31/2005
Source: Nasdaq.com

% of index 7.29 5.86 4.08 3.82 3.33 3.20 3.01 2.80 2.75 2.59 38.73

Top 10 industry groups 1 Information technology hardware 2 Computer software/services 3 Other services 4 Retail/wholesale trade 5 Pharmaceuticals & biotechnology 6 Telecommunications 7 Manufacturing 8 Health 9 Consumer goods 10 Transportation Total

% of index 35.09 20.25 11.65 11.54 9.31 4.50 3.13 2.67 1.06 0.80 ging, the S&P will catch up. However, if the Nasdaq is headed down, the S&P 500 won’t be far behind, which creates an opportunity for SPY and S&P futures swing traders.

Using momentum to define the QQQQ/SPY trend

FIGURE 2 THE RELATIVE STRENGTH INDEX (RSI) During the uptrend, the RSI peaked well above 60 and bottomed at 40 or higher. The indicator oscillated between 60 and 40 during the NovemberDecember trading range. In February, the RSI peaked well below 60, and the downtrend was confirmed when the RSI closed below 40 in March.
Nasdaq 100 index tracking stock (QQQQ) 39.00 38.00 37.00 36.00 35.00 34.00 33.00 32.00 31.00 30.00 RSI 60 40 Aug. 2003 Sept. Source: CQGNet Oct. Nov. Dec. 2004 Feb. March

Finding a way to define uptrend or downtrend conditions is always a challenge. The sim100.00 plest technique is to use a moving average: If the QQQQ/SPY ratio closes above the moving average, the trend is up, and if the ratio closes below the moving average, the trend is down. A problem with this approach is choosing the look-back period. A tooshort moving average will result in many false signals; a too-long moving average will increase the lag time between a change in direction in the ratio and the ratio-moving average crossover signal. Another approach is to determine the trend using a momentum calculation — in this case, a 14-bar relative strength index (RSI). Up momentum indicates a rising trend, while down momentum implies a falling market. Although there is a problem of selecting appropriate levels to define momentum as negative or positive, momentum behaves differently during rising trends vs. falling trends. This behavior provides the basis for the trend-determination rules. (See “Key concepts and definitions” for more information about the RSI.) During an uptrend, momentum-indicator readings are shifted upward — the indicator lows and highs are both higher than they would be during a period when price is moving sideways. The indicator does not reach the typical oversold level before the upward trend reasserts itself following a countertrend

53

www.activetradermag.com • September 2005 • ACTIVE TRADER

FIGURE 3 QQQQ/SPY RATIO WITH RSI At point A the RSI peaked above 70 and, as the ratio began to decline, the RSI dropped below 50. Turning up twice from above 40 (points B and C) indicated the ratio could turn up again. At point D, the RSI closed below 40, signaling the ratio was now in a bearish momentum trend.
Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts (QQQQ/SPY), daily 0.335

0.330

0.325

A

RSI 60

B
November 1 8 15 22 December 1 6 13 20

C
27

40

D
2005 3 10 18

Source: CQGNet

FIGURE 4 DEFINING THE TREND At point A, the RSI was below 40, indicating bearish momentum. At point B, it had crossed above and back below 50 — continuing to indicate a bearish trend for the ratio. The RSI bottomed above 40 at point C, indicating a trading range. At Point D, the RSI peaked just above 60 and indicator readings were shifting upward, a bullish sign for the market.
Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts (QQQQ/SPY), daily 0.320

0.315

0.310

0.305 RSI

D B C
August 2 9 Source: CQGNet

60

40 October 20 27 1 11

A
16 23

September 1 7 13

move. Similarly, during a downtrend, momentum-indicator readings are shifted lower and the indicator fails to reach the typical overbought levels during countertrend rallies. This requires adjusting the momentum indicator’s overbought and oversold levels accordingly for the trending period. Figure 2 shows daily QQQQ bars with a 14-bar RSI. The RSI has two horizontal lines at 60 and 40. Throughout the uptrend from August 2003 to January 2004, the RSI peaked above 60 and bottomed at 40 or higher. Then the market began to turn down in February and March of 2004. At first the RSI didn’t fall below 40, but when the market rallied in early February, it peaked well below 60. Finally, in early March the RSI closed below 40. The 60 and 40 readings will be the key RSI thresholds in this analysis. During an uptrend, peak readings should be above 60 and low readings should hold above the oversold level of 40. If the trend is down, the RSI will tend to peak near 60 and make lows below 40. If the RSI begins fluctuating between 60 and 40, the market is moving into a trading range. Figure 3 shows the QQQQ/SPY ratio with the RSI. It contains an example of the RSI switching from indicating bullish momentum to bearish momentum. During early December 2004, the RSI peaked well above 60 and even above 70 (point A) and as the ratio began to work its way lower, the RSI dropped below 50, but turned up from above 40 twice (points B and C). At point D, the RSI closed below 40, signaling the ratio was now in a bearish trend. (There was a warning the ratio was turning bearish when the RSI closed below the rising trendline connecting points B and C one day before the RSI closed below 40.) Figure 4 shows an example of the RSI indicating a downtrend in the QQQQ/SPY ratio and then changing to

54

www.activetradermag.com • September 2005 • ACTIVE TRADER

indicate a bullish trend. At point A, the RSI is below 40, indicating bearish momentum. At point B, the RSI has crossed above 50 (but well below 60) and back below 50, so it continues to indicate bearish momentum. At point C, the RSI troughs at 40, indicating a trading range. At point D, it peaks just above 60 at 60.15 — not a definitive bullish signal, but this reading is higher than point B, so the momentum readings are shifting upwards, which is a bullish sign for the market as the QQQQ is starting to lead.

FIGURE 5 UPTHRUST TRADE The false breakout was confirmed (upthrust) when the market closed back below the December high. The short sale was signaled on the close, and the risk point is above the high of the entry bar. Another upthrust occurred in April, which preceded another new low in the downtrend.
S&P 500 Depositary Receipts (SPY), daily Upthrust 122.00

120.00 Short 118.00

A swing trading approach: Springs and upthrusts
A swing trading method needs two components: a definition of the market trend and an entry criteria for trading in the direction of the trend. Because QQQQ tends to lead SPY, the following strategy trades SPY based on the trend of the QQQQ/SPY ratio. The entry rules are based on two Richard Wyckoff setups called upthrusts and springs (see “Related reading” ). An upthrust is a breakout and close above a previous resistance level that fails. The breakout usually leads to an overbought situation, and if the false breakout is against the direction of the trend, the subsequent decline can be dramatic. A spring is a false breakdown and close below support. Once support has been broken, pessimism rises and the market becomes oversold and then rallies. An example of an upthrust occurred in March 2005. Figure 5 shows SPY from December 2004 through April 2005, along with the QQQQ/SPY ratio and the 14-day RSI of the QQQQ/SPY ratio. The low RSI readings reflect a clear downtrend. On March 4, SPY made a new high closing price for the year. This was an employment-report Friday, and the news was good. Here’s commentary from Briefing.com after the close of trad-

Short

116.00

114.00

Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts (QQQQ/SPY), daily

0.33 0.32 0.31

RSI

60 40

20 27

2005 3 10 18 24

Downtrend February March 1 7 14 22 1 7

April 14 21 28 1 11 18

20

Source: CQGNet

ing that day: “Stronger than expected job creation, absent worries of inflation, ignited a broadbased rally that kept a bullish bias intact and closed the blue chip indices at 3-year highs. [A]ll three major Dow Indices hit new 52week highs, as the Dow Industrials surpassed the 10,900 level for the first time since June 2001 with gains realized in 27 of 30 components. [T]he S&P 500 erased 2005 losses while the S&P 400 MidCap (+1.1 per-

cent) and the S&P 600 Small Cap 600 Index (+1.2 percent) have both touched all-time highs, while the Russell 2000 Index (+1.0 percent) also surged but remains just shy of a new 2005 high, as the index touched 654.27 on Jan. 3. [T]he Nasdaq, however, despite a strong performance, is still off roughly 4.8 percent in 2005.” Here we have a situation where the market is digesting great fundamental

ACTIVE TRADER • September 2005 • www.activetradermag.com

55

FIGURE 6 NOT JUST THE RATIO From late January through early March, SPY led an uptrend against the direction of the QQQQ/SPY ratio. Don’t short SPY simply because it is climbing and the ratio is falling.
S&P 500 Depositary Receipts (SPY), daily 123.00 122.00 Out 121.00 120.00 Short Out Short 119.00 118.00 117.00

close (with the risk point above the entry bar). Four days later, SPY made a new low in the downtrend.

Caveat
The QQQQ/SPY spread is a good barometer of the broader trend, but it should not be used as the sole indicator of the market. Sometimes SPY will trend higher while the ratio is declining, as occurred between late January and early March. You would not want to simply short SPY because it was climbing and the QQQQ/SPY ratio was falling. There were, as Figure 6 shows, two additional trades based on the upthrust concept during this rally — one win and one loss — prior to the setup at the March high.

Springs
Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts (QQQQ/SPY), daily 0.32

0.31

RSI

60 40

18 24 Source: CQGNet

February 1 7

14

22

March 1 7

20 14

news and key market indices are making new highs. However, if the market was actually strong, QQQQ should have been leading the pace, not lagging so significantly — capital should have been moving into technology, not out of it. These conditions set up a short sale in SPY or the S&P 500 futures. When the market closed back below the December high, the false breakout was confirmed and a short sale was signaled on the close. The stop point would be above the high over the entry bar. A logical place to take profits was after the
56

market took out the support level defined by the late-February low, as selling would likely accelerate, creating good liquidity for covering short positions. A second, shorter-term trade example occurred in April 2005. The RSI of the QQQQ/SPY ratio was still indicating a (weak) downtrend was in place by climbing above 50 only twice and reversing immediately each time. On April 7, SPY closed above the previous week’s high, then closed back below it, signaling a short sale on the

Figure 7 is an example of the QQQQ leading the way and trading SPY from the long side. The RSI of the QQQQ/SPY ratio began generating bullish readings with a close at 61 on Sept. 20. Later in the month, the RSI dropped below 50 for just one day and climbed back above that level, providing further bullish evidence. On Oct. 20 SPY dropped through the September support level and closed back above it, completing a spring setup and signaling a long trade on the close. Three days later, the market traded below the entry bar’s low, so the trade was stopped out. But the market reversed again, closing near the high of the day and back above the entry bar’s low. The situation qualifies as a new spring because the close (A) is above a support level, even though the support level is three days away. However, you could enter the market after it trades above the bar’s high (B) or wait for a bar that closes back above the September support level (C). A move above the early October high is a logical profit-taking level, as you would be selling into a very optimistic

www.activetradermag.com • September 2005 • ACTIVE TRADER

FIGURE 7 SPRING SETUP This “spring” setup triggered a trade on the close above support (A), with risk below the entry bar. The first entry setup can fail, as was the case here. Then, use the second spring to buy or place a buy stop over the high of the entry rule bar or wait for the close back above the original support level.
S&P 500 Depositary Receipts (SPY), daily

116.00

market.

Spreading the wealth
114.00

112.00

C
Long Out Nasdaq 100 index tracking stock/S&P 500 Depositary Receipts (QQQQ/SPY), daily 0.32

B A

110.00

RSI

0.31 80 60 40

Using relative strength to gauge for fund flows removes the challenge of trying to understand the implications of every tick of the market. The QQQQ/SPY ratio is a good tool for identifying the flow of funds between technology and more defensive largecap stocks, such as financials and health care. Traders can benefit from this information by incorporating trading signals that exploit price patterns that can form when the market is reversing direction. The upthrust and spring patterns illustrated here are good examples. These patterns take advantage of a market’s tendency to make a false breakout of support or resistance, which can precede strong price movement in the direction of the longer-term trend.

October 13 20 Source: CQGNet 27 1 11 18 25

November 1 8

Related reading
“Trading the Wyckoff way: Buying springs and selling upthrusts,” by Henry O. (Hank) Pruden (Active Trader, August 2000). Understanding Wyckoff’s springs and upthrusts can help you identify low-risk, high-reward trade setups based on false breakouts. “Wyckoff axioms jumps and backups,” by Hank Pruden (Active Trader, JanuaryFebruary 2001). Not every breakout will turn into a good trade, but using Wyckoff’s concepts of “jumps” and “backups” can help you sort the good trades from the bad. “The telltale spread,” by Thom Hartle (Active Trader, November 2003). Analyzing the relationship between the Emini Nasdaq 100 and the E-mini S&P 500 can indicate when the broad market is making a genuine move or when it’s faking people out. “RS System No. 1,” by Thomas Stridsman (Active Trader, October 2000). This is a short-term relative-strength (RS) system that compares the stock price with its underlying market index, but also looks for confirmation from volume. “Relative strength bands,” by Thomas Stridsman (Active Trader, March 2001). This system uses a mix of relative strength (not RSI) analysis and Bollinger Bands to identify markets that are about to break out of congestion areas. “Finding strength with the ratio spread,” by Thomas Stridsman (Active Trader, August 2000). The ratio spread provides a better alternative to moving averages for determining trend and support and resistance. “Indicator Insight: Relative Strength,” by Active Trader Staff (Active Trader, June 2002). Relative strength can be measured and interpreted in several different ways. “Indicator Insight: Relative Strength Index,” by Active Trader Staff (Active Trader, August 2001). The relative strength index (RSI) is a momentum oscillator used to identify short-term momentum extremes (overbought and oversold points). J. Welles Wilder, developer of the RSI, provides step-by-step instructions for calculating and interpreting the indicator. Special notice: From Aug. 1 to 31, these articles will be available for 30 percent off through the Active Trader online store (www.activetradermag.com/purchase_art icles.htm). You can download them directly to your computer as PDF files for easy viewing and printing.

57

ACTIVE TRADER • September 2005 • www.activetradermag.com

CURRENCY BASICS

Comparing moving averages
This review provides simple ways to interpret and apply the most commonly used moving averages.
BY THOM HARTLE

T

FIGURE 1 — SIMPLE MOVING AVERAGE

he moving average is a fundaThe 10-bar SMA smoothes price but lags market turning points. When price slips (point B), the moving average continues to rise. But at point mental tool for smoothing price C, the SMA turns down, following the declining prices. action and revealing trend. In the early days of technical analysis, traders simply looked for a crossover of the moving average (i.e., when price crosses above or below its average) as a sign the trend had changed. Today, traders have shorter-term trading horizons, so the moving average is also used as a “decision filter.” For example, if price is above its moving average, then the trend is up, and the trader will look for short-term buy setups (rather than sell setups). Although moving averages can simplify price action and highlight the underlying trend, they also trail behind (lag) price action — the longer the moving average, the more it lags price. There are several types of moving averages; traders originally used the simple moving Source: eSignal average (SMA) because it was easy to understand and calculate. Over time, other smoothing techniques over time (e.g., the 50-day and 200-day moving averages), were incorporated to reduce the lag inherent in the SMA. but there is no moving average length that will best reflect Here, we will compare the three most common moving the trend in all conditions and markets. The following average types — simple, weighted, and exponential. examples use several representative moving average There is no “correct” number of bars to use in a moving lengths. average. Certain “look-back” periods have become popular

There is no “correct” number of bars to use in a moving average. Certain “look-back” periods have become popular over time, but there is no moving average length that will best reflect the trend in all conditions and markets.
58 November 2005 • CURRENCY TRADER

FIGURE 2 — 10-DAY SMA At point A, the Euro/U.S. dollar peaks, and its 10-day SMA turns down three days later. The retracement (point B) doesn’t affect the SMA much, and price rallies above its moving average well before the SMA turns up (point C).

Simple moving average
Our first example uses a 10-bar SMA, but technicians commonly use moving averages as long as 200 days (approximately one year in trading days) to define the long-term trend, as well as a combination of moving averages such as four-, nine-, and 21-day periods to simultaneously track three trends. To calculate a 10-bar SMA, add the last 10 bars’ closes and divide by 10. When the next bar closes, the most distant closing price in the look-back period is dropped, the new bar’s closing value is added, and the new sum is divided by 10. Figure 1 shows simulated prices with a 10-bar SMA. (Notice it takes 10 closes to plot the first SMA value.) At point A, the simple moving average is near the middle of the price range. As price begins to rise, the simple moving average turns up a few closes later, illustrating the moving average’s inherent lag. The average continues to rise despite a short-term retracement (point B), and both values climb after that. Next, price forms a double top and, as it declines, the simple moving average crests and turns down (point C). During the drop, the SMA trended down despite a final countertrend rise. Figure 1 shows the benefits and drawbacks of simple moving averages. Despite short-term, counter-trend movements, the SMA filters out this noise and continues in the direction of the trend. But the SMA’s lag is noticeable when price changes direction. Figure 2’s daily chart of the Euro/ U.S. dollar currency pair shows the same characteristics as in Figure 1. Here, the simple moving average follows the market as it peaks and then turns down (point A). The market

Source: eSignal

FIGURE 3 — 10-DAY WMA VS. SMA The weighted moving average tracks price action closer than the simple moving average. The WMA’s turning points (points A and C) occur before the SMA’s. Point B’s price rise did not materially affect the WMA.

Source: eSignal

CURRENCY TRADER • November 2005

59

CURRENCY BASICS continued

FIGURE 4 — 10-DAY EMA VS. SMA The EMA turns each time price closes above or below it (points A and B), and the EMA turns up slightly at point C.

point B’s countertrend move did not derail the descent of the WMA.

Exponential moving averages
The exponential moving average (EMA) uses a special algorithm (the “smoothing constant”) that weights price by a percentage factor. There are two versions of the EMA calculation. Both versions use a smoothing constant to weight the closing price and adjust the previous day’s EMA value. The constant uses a formula to approximate the value of a SMA: SC = 2/(n + 1) where n = the look-back period for a simple moving average SC = a smoothing constant between 1 and zero

Source: eSignal

Therefore, if n = 10: retraces (point B), but the moving average continues to trend lower. At point C, price rallies above its moving average and the MA turns up. We see the same lag at points A and C when the market reverses the trend. SC = 2/(10+1) = 2/11 = 0.1818 The first version of the EMA calculation multiplies today’s close by the smoothing constant and yesterday’s EMA by 1 minus the smoothing constant: (today’s close * 0.1818) + (yesterday’s EMA * 0.8182) The second approach uses the following formula: (today’s close - yesterday’s EMA)* 0.1818) + yesterday’s EMA The formulas show how important today’s close is relative to the previous day’s EMA. If today’s close is above yesterday’s EMA for the first time, the EMA will immediately turn up. If today’s close is less than yesterday’s EMA, the average will immediately turn down. There is no lag between today’s close and the direction of the EMA. During trading ranges, however, the EMA will flip back and forth. Figure 4 compares a 10-day EMA to the 10-day SMA. At points A and D the EMA reverses direction once the Euro crosses above or below it. It is a little less noticeable, but at
November 2005 • CURRENCY TRADER

Weighted moving averages
One reason the SMA lags price is because all prices are equally weighted — the current close and the first close have the same impact on the average. A weighted moving average (WMA) uses specific multipliers to weight each price, giving the most weight to the most recent price and reducing this emphasis the further back in time you go. Thus, the most current price impacts the WMA more than the last price in the look-back window. For example, a 10-bar WMA multiplies the current price by 10, the next most recent price by 9, the next by 8, and so on. The sum of these weighted prices is divided the sum of the weights, which is 55 (10 + 9 + 8 + 7… 1) in this example. Figure 3 compares a 10-bar WMA (blue) to the 10-bar SMA (red). Notice the WMA crests and troughs (points A and C) ahead of the SMA. The WMA responds much quicker to the reversal patterns at the top and bottom, which took approximately five days. In contrast, the SMA required either more time or a more dramatic price move to reverse its trend. But
60

FIGURE 5 — WMA/EMA CROSSOVER SIGNALS If the 20-day EMA is rising, the trend is up and price moves below the threeday WMA represent buy opportunities (green area). If the EMA is falling, the area between the EMA and WMA can be considered a sell zone (red).

points B and C, the EMA turns up because price closes above it. The EMA is best applied to markets that are prone to spike reversals. A market that moves into a lengthy trading range will be constantly crisscrossing this average.

Applying moving averages
Think of the moving average as exactly what its name says: the average price. Many years ago, traders viewed a cross of the moving average as a sign the trend has changed direction. But it doesn’t make sense the trend has changed direction simply because price is somewhere above or below its average. On the other hand, if the average price is rising or falling, then it provides information about the trend’s direction. A rising average price indicates the trend is up; a falling moving average reflects a descending trend. These two ideas can be used together with two different moving averages and look-back periods. Figure 5 shows the Euro/ U.S. dollar with a 20-day EMA and a three-day WMA. For example, you could consider the direction of the exponential moving average as the trend. If the EMA is rising, the trend is up. If the EMA is falling, the trend is down. Next, the three-day WMA can be considered a level of value. If the trend is up as defined by a rising EMA and price is below yesterday’s WMA (we don’t know today’s reading until the close), then price is cheap relative to the upward trend. Consider the area between the rising EMA and the WMA as a buy zone (green area) for intraday and short-term setups. If the trend is down (the EMA is declining) and current price is above yesterday’s WMA, then price is expensive relative to the downward trend. The area between the falling EMA and the WMA is a sell zone (red area). Moving averages can be used as a frame of reference to indicate the trend and whether the current price is cheap or expensive relative to the trend. Considering how trends in forex tend to persist, having a tool that keeps you looking for opportunities on the right side of the trend is a very wise first step towards profitability.

Source: eSignal

Related reading
“The weighted moving average system” Active Trader, November 2004. A test of a WMA crossover system on a futures portfolio. “Trend vs. countertrend indicators” Active Trader, June 2004. An explanation of how technical indicators work and the differences between trend-following and countertrend calculations. “Indicator Insight: Weighted and exponential moving averages” Currency Trader, January 2005. How to calculate and interpret weighted and exponential moving averages.Includes simple comparison tests of different moving averages. “Indicator Insight: Simple moving average,” Currency Trader, December 2004. A primer on the calculation and application of simple moving averages. You can purchase and download past articles at www.activetradermag.com/purchase_articles.htm.

CURRENCY TRADER • November 2005

61

TRADING Strategies

The ETF money trail
Quarter-by-quarter analysis shows how tracking ETF performance can tip you off to overall market performance and trade opportunities in certain sectors.

62

www.activetradermag.com • December 2005 • ACTIVE TRADER

TABLE 1 TOP 50 EXCHANGE TRADED FUNDS The Top 50 Exchange Traded Funds based on ranking of the 10-day average volume on June 30, 2005. Name Merrill Lynch Biotech HOLDRS DIAMONDS Trust Series I iShares Dow Jones Select Dividend Index Fund iShares MSCI Emerging Index Fund iShares MSCI EAFE Index Fund iShares MSCI Germany Index Fund iShares MSCI Hong Kong Index Fund iShares MSCI Japan Index Fund iShares MSCI Malaysia Index Fund iShares MSCI Taiwan Index Fund iShares MSCI Mexico Index Fund iShares MSCI South Korea Index Fund iShares MSCI Brazil Index Fund Merrill Lynch Internet HOLDRS iShares NASDAQ Biotechnology Index Fund iShares Lehman 7-10 Year Treasury Bond Fund iShares Goldman Sachs Software Index Fund iShares S&P MidCap 400 Index Fund iShares S&P MidCap 400/Barra Value Index Fund iShares S&P SmallCap 600 Index Fund iShares S&P 500/BARRA Value Index Fund iShares S&P 500 Index Fund iShares Russell 1000 Value Index Fund iShares Russell 1000 Growth Index Fund iShares Russell 2000 Index Fund iShares Russell 2000 Value Index Fund iShares Russell 2000 Growth Index Fund iShares Dow Jones U.S. Real Estate Index Fund MidCap SPDR Trust Series I iShares S&P 100 Index Fund Merrill Lynch Market Oil Service HOLDRS Merrill Lynch Pharmaceutical HOLDRS NASDAQ-100 Index Tracking Stock Merrill Lynch Regional Bank HOLDRS Merrill Lynch Retail HOLDRS iShares Lehman 1-3 Year Treasury Bond Fund Merrill Lynch Semiconductor HOLDRS SPDR Trust Series I iShares Lehman 20 Year Treasury Bond Fund Merrill Lynch Telecom HOLDRS Merrill Lynch Utilities HOLDRS Select Sector SPDR Fund - Basic Industries Select Sector SPDR Fund - Energy Select Sector Select Sector SPDR Fund - Financial Select Sector SPDR Fund - Industrial Select Sector SPDR Fund - Technology Select Sector SPDR Fund - Consumer Staples Select Sector SPDR Fund - Utilities Select Sector SPDR Fund - Health Care Select Sector SPDR Fund - Consumer Discretionary
Source: eSignal

BY THOM HARTLE

irst launched in the early 90s, Exchange Traded Funds (ETFs) have continued to proliferate on Wall Street. Today, there are more than 170 ETFs listed for trading — an increase of more than 50 in less than two years. These instruments, which trade like regular stocks, represent everything from major stock indices to specific market sectors and instrument groups (bonds, etc.). The appeal to investors and traders is twofold. First, people can buy assets that closely replicate the returns of popular market indices fund managers use as benchmarks. Second, traders can use sector ETFs to capitalize on specific market areas that are outperforming the broad market. “Following the money: Quarterly momentum and ETFs” (Active Trader, June 2002) detailed a strategy for identifying ETF leaders early in the quarter. The premise was that money managers will commit capital to stocks offering the best opportunity for profit for the quarter. Thus, the ETFs leading the market at the beginning of a quarter had the potential to maintain their dominance for the entire quarter. The following analysis reviews the findings of that article and, using the same procedures, explores how its concepts have held up in the eighteen months since publication.

F

Narrowing the ETF field
Because there are so many ETFs to consider, the first step is to narrow the field to a reasonable number by focusing on

Symbol BBH DIA DVY EEM EFA EWG EWH EWJ EWM EWT EWW EWY EWZ HHH IBB IEF IGV IJH IJJ IJR IVE IVV IWD IWF IWM IWN IWO IYR MDY OEF OIH PPH QQQQ RKH RTH SHY SMH SPY TLT TTH UTH XLB XLE XLF XLI XLK XLP XLU XLV XLY

ACTIVE TRADER • December 2005 • www.activetradermag.com

63

TABLE 2 TOP 10 TRADED ETFS BASED ON AVERAGE VOLUME These ETFs were the most actively traded during the review period. Name NASDAQ-100 Index Tracking Stock SPDR Trust Series I Merrill Lynch Semiconductor HOLDRS iShares Russell 2000 Index Fund Select Sector SPDR Fund - Energy Select Sector iShares MSCI Japan Index Fund DIAMONDS Trust Series I Merrill Lynch Market Oil Service HOLDRS Select Sector SPDR Fund - Financial MidCap SPDR Trust Series I
Source: eSignal

Symbol 10-day avg. vol. QQQQ SPY SMH IWM XLE EWJ DIA OIH XLF MDY 88,389,774 48,865,710 21,776,440 19,471,120 14,061,760 7,610,670 7,073,800 5,494,090 3,528,900 2,461,530

fourth, reflecting the resurgence of small-cap stocks since the market bottom in late 2002 (IWM has risen over 100 percent since that low). Not surprisingly given the recent oil surge, two energy-oriented ETFs are in the top 10: the Select Sector SPDREnergy (XLE) and the Merrill Lynch Oil Service HOLDRS (OIH). Table 3 shows ETFs from “Following the money” that didn’t meet the current average daily volume criteria, and the ETFs that replaced them for the current study.

ETF recap
The first step in the analysis is to calculate the performance of the 50 top-volume ETFs for the beginning of every quarter. Returns were measured from the end of the previous quarter to the second Friday in the new quarter. (The second Friday was chosen to ensure market reaction to the all-important monthly employment number.) Then the end-ofquarter returns were calculated for the top-five ETFs (as well as for QQQQ and SPY) to see if there was any connection to early-quarter performance.

the most-liquid instruments — in this case, the 50 ETFs with the highest 10-day average volume (Table 1). Table 2 shows the top 10 from this list sorted by volume. The two top-traded ETFs are the Nasdaq-100 Index Tracking TABLE 3 NEW VS. OLD

Stock (QQQQ) and the Standard & Poor’s Depositary Receipts (SPY), which averaged more than 80 million and 48 million shares daily, respectively, for the last 10 days of June 2005. The iShares Russell 2000 Index Fund (IWM) is

The first column lists the ETFs from the “Following the money” article that did not make the volume cut for this analysis. The second column shows the ETFs that replaced the deleted ETFs. ETF (Deleted) Fresco DJ EURO STOXX 50 iShares MSCI Austria Index Fund iShares MSCI EMU Index Fund iShares MSCI United Kingdom Index Fund iShares Russell 1000 Index Fund iShares Russell 3000 Index Fund iShares S&P 500 Index Fund Merrill Lynch B2B Internet HOLDRS Merrill Lynch Internet Infrastructure HOLDRS Merrill Lynch Software HOLDRS Vanguard Total Stock Market VIPERs
Source: eSignal

Symbol FEZ EWO EZU EWU IWB IWV IVW BHH IIH SWH VTI

ETF (Added) iShares Dow Jones Select Dividend Index Fund iShares Goldman Sachs Software Index Fund iShares MSCI Germany Index Fund iShares MSCI Mexico Index Fund iShares Russell 2000 Value Index Fund iShares S&P 100 Index Fund iShares S&P MidCap 400 Index Fund iShares S&P MidCap 400/Barra Value Index Fund iShares S&P SmallCap 600 Index Fund Merrill Lynch Regional Bank HOLDRS Merrill Lynch Telecom HOLDRS

Symbol DVY IGV EWG EWW IWN OEF IJH IJJ IJR RKH TTH

64

www.activetradermag.com • December 2005 • ACTIVE TRADER

“Following the money” showed the early-quarter leaders often placed fairly high at the end of the quarter — as long as there was broad-based buying of ETFs and confirmation from SPY and QQQQ at the start of the quarter. For example, if only 20 percent of ETFs were up by the second Friday of the quarter, chances were high the quarter would perform poorly overall, and even those ETFs showing positive returns at the beginning of the quarter would likely be dragged down by the broader market’s downtrend. On the other hand, if 80 percent of ETFs were up at the end of the first two weeks of the quarter and QQQQ and SPY were also strong during this period, the odds were better that the best-performing ETFs at the beginning of the quarter would also be among the strongest at the end of the quarter. The ETF performance in the six quarters since “Following the money” was published are summarized in Table 4 (2004) and Table 5 (the first two quarters of 2005). The tables show the returns for each of the topfive ETFs (along with QQQQ and SPY), as well as the average return for these five ETFs on the second Friday of each quarter; the same statistics are shown for the end of the quarter. Also included for each interval are the percent ages of all 50 top-volume ETFs that were positive and their median return.

TABLE 4 QUARTERLY BREAKDOWN OF THE ETF PERFORMANCE FOR 2004 The first column for each quarter lists the top-five ETFs after two weeks (along with QQQQ and SPY). In addition, the average return of these top five ETFs (excluding the QQQQ and SPY) is listed. Below, the percentage of all 50 top-volume ETFs with positive returns (and their median return) is shown. The next column lists the individual returns for the initial top-five ETFs at the end of the quarter, their rank, and their average return. Finally, the percentage of all 50 ETFs up for the quarter and their median endof-quarter return are shown. First Quarter (2004) First End of two weeks quarter ETF SMH EWZ EWW EWT EWH QQQQ SPY Average % Bullish Median 1/9/04 Rank 3/31/04 Rank 7.86% 1 -4.80% 50 7.65% 2 -1.18% 44 7.14% 3 21.55% 1 6.32% 4 8.36% 7 5.30% 5 6.00% 14 3.48% 14 -1.70% 46 1.00% 33 1.64% 32 6.85% 5.99% 84.00% 1.42% 80.00% 3.66% ETF HHH EWY SMH EWT QQQQ BBH SPY Average % Bullish Median Second Quarter First End of two weeks quarter 4/8/04 Rank 6/30/04 Rank 9.57% 1 26.48% 1 5.29% 2 -13.66% 50 3.92% 3 -4.23% 39 3.28% 4 -11.17% 47 3.07% 5 5.30% 4 3.07% 6 3.06% 7 1.12% 24 1.26% 14 5.03% 0.10% 76.00% 1.10% 56.00% 0.44%

Third Quarter First End of two weeks quarter ETF EWM XLE IYR EWH TLT QQQQ SPY Average % Bullish Median
Source: eSignal

Fourth Quarter First End of two weeks quarter ETF EWZ EWY XLE IGV EWT QQQQ SPY Average % Bullish Median 10/8/04 Rank 12/31/04 Rank 4.64% 1 19.96% 3 4.60% 2 17.00% 7 3.61% 3 3.98% 45 3.40% 4 23.05% 1 3.36% 5 12.71% 16 1.25% 22 13.59% 11 0.67% 27 8.15% 31 3.92% 15.34% 78.00% 0.89% 96.00% 9.54%

7/9/04 Rank 9/30/04 Rank 4.96% 1 4.19% 12 1.43% 2 10.96% 3 1.31% 3 6.86% 5 1.18% 4 9.77% 4 0.87% 5 5.50% 9 -5.27% 43 -6.89% 44 -2.44% 27 1.26% 34 1.95% 7.46% 18.00% -2.36% 40.00% -1.11%

2004 performance
In Q1 2004, 84 percent of the 50 top-volume ETFs were up on the second Friday; 80 percent were up at the end of the quarter. Interestingly, the Semiconductor

HOLDRS (SMH), which was the topperforming ETF at the start of the quarter (+7.86 percent), was at the bottom of the list at the end of the quarter (-4.80 percent). Still, three of the top five ETFs at the beginning of the quarter had positive returns at the end of the month and

also outperformed QQQQ and SPY. The end-of-quarter average return for the top five was 5.99 percent, mostly because the iShares MSCI Mexico Index Fund (EWW), which started ranked No. 3 and finished No. 1, climbed 21.55 percent. Two of the top five, EWW and
65

ACTIVE TRADER • December 2005 • www.activetradermag.com

TABLE 5 THE FIRST TWO QUARTERS OF 2005

EWT, were in the top 10 at the end of the quarter. Many of the ETFs were The same statistics from Table 4 are shown here for 2005. experiencing new highs for the year heading into March First Quarter Second Quarter 2004, but a correction unfoldFirst End of First End of ed that erased much of the two weeks quarter two weeks quarter first quarter’s gains. ETF 1/14/05 Rank 3/31/05 Rank ETF 4/8/05 Rank 6/30/05 Rank Consequently, the average EWM 4.76% 1 -5.03% 37 EWY 3.74% 1 0.60% 37 return for the top five at the end of the quarter was posiEWY 2.05% 2 8.75% 3 IBB 2.86% 2 6.85% 9 tive (5.99 percent), but below TLT 1.83% 3 0.88% 8 EEM 2.71% 3 5.92% 11 the second Friday’s average OIH 1.70% 4 13.02% 2 BBH 2.60% 4 18.83% 1 return (6.85 percent). The XLP 0.48% 5 -0.26% 10 XLV 2.01% 5 3.92% 19 median return on the second Friday for all 50 ETFs was QQQQ -3.72% 34 -8.38% 46 QQQQ 0.19% 18 0.57% 38 1.42 percent; at the end of SPY -2.18% 17 -2.41% 25 SPY 0.03% 22 1.03% 35 quarter it was 3.66 percent. Average 2.16% 3.47% Average 2.79% 7.22% Figure 1 shows the first two-week and end-of-quar% Bullish 14.00% 18.00% % Bullish 44.00% 78.00% ter periods sorted by the highest returns for the secMedian -2.83% -2.46% Median -0.13% 2.69% ond Friday of the quarter. Many of the ETFs did well Source: eSignal for the quarter compared to the top five. The start of the second quarter looked FIGURE 1 BEGINNING VS. END OF FIRST QUARTER 2004 positive, with 76 percent of all 50 ETFs The ETFs are ranked from the highest to lowest return on the second Friday showing positive early returns and of the quarter (green bars), while the red bars represent the return for the QQQQ making it into the top five. The same ETFs at the end of the quarter. Looking to the right, many of the ETFs median return for all 50 ETFs was 1.10 did much better than their initial returns. percent on the second Friday, but only 56 percent of the ETFs showed a positive return by the end of the second quarter, with a final median return of just 0.44 percent. The average return for the top five was only 0.10 percent for the quarter. Still, two of the initial top five (excluding the QQQQ) were in the top 10 at the end of the second quarter, and the Internet HOLDRS (HHH) were No. 1 at both the start and the finish. The SPY was nearly unchanged from the second Friday to the end of the quarter. Figure 2 shows how mixed the returns were at the end of the quarter. At the beginning of Q3 2004 it looked like the poor performance at the end of Q2 was carrying over. Only 18 percent of the ETFs had positive returns by the second Friday, with a median return of -2.36 percent for all 50 ETFs. QQQQ was Source: eSignal down 5.27 percent and SPY was down
www.activetradermag.com • December 2005 • ACTIVE TRADER

66

2.44 percent. However, by the end of the quarter there were signs of recovery: Forty percent of all 50 ETFs were positive, with a median return of -1.11 percent. QQQQ continued to drop, falling to -6.89 percent at the end of the quarter, but SPY recovered (-2.44 percent vs. +1.26 percent). The initial top-five ETFs held up well in Q3. The average return at the end of the quarter was 7.46 percent, 5.51 percent above the quarter’s starting point. Also, four of the original top five were in the top 10 at the end of the quarter. Figure 3 shows how the early leaders did well for the quarter, while the majority of ETFs declined. As in “Following the money,” the fourth quarter of 2004 showed the best returns for the year. At the start, 78 percent of the ETFs were positive and 96 percent were up by the end of the quarter. The top five averaged 15.34 percent for the quarter (adding almost 12 percent from the second Friday’s average return), nearly outperforming QQQQ’s return (up 12.34 percent from the second Friday to 13.59 percent for the quarter) and nearly twice SPY’s return (8.15 percent). At the start of the quarter, the median return for all 50 ETFs was 0.89 percent; it jumped to 9.54 percent by the end of the quarter. Figure 4 compares fourthquarter 2004 performance. The early returns were below +5 percent, but the buying was broad based. Also, three of the initial top five placed in the top 10 at the end of the quarter.

FIGURE 2 BEGINNING VS. END OF SECOND QUARTER 2004 The top-performing ETF (HHH) at the start of the quarter also was the topperforming ETF at the end.

Source: eSignal

FIGURE 3 BEGINNING VS. END OF THIRD QUARTER 2004 The early leaders (left-hand side) continued to advance, while the remainder of ETFs declined for the quarter.

2005 performance
The first quarter of 2005 got off to a rocky start, with just 14 percent of the 50 top-volume ETFs showing positive returns (-2.83 percent median return), and QQQQ down 3.72 percent and SPY down 2.18 percent. Only 18 percent of the top-50 ETFs were positive at the end of the quarter (-2.46 percent median return). However, the initial top-five ETFs had an average return of 2.16 percent in the first two weeks and four of these

Source: eSignal

www.activetradermag.com • December 2005 • ACTIVE TRADER

67

FIGURE 4 BEGINNING VS. END OF FOURTH QUARTER 2004 ETFs enjoyed a dramatic rise for the quarter, as just two failed to show positive returns at the end of the quarter.

Source: eSignal

FIGURE 5 BEGINNING VS. END OF FIRST QUARTER 2005 Despite the overall weak reading of positive ETFs at the start of the quarter (just 14-percent with positive returns and median return of -2.83), three of the initial top five were positive at the end of the quarter.

were still in the top 10 at the end of the quarter. Figure 5 shows the two sets of returns sorted by initial review period. At the start of the second quarter of 2005, only 44 percent of the top-50 ETFs were up, but by the end of the quarter 78 percent were positive. The first two weeks had a median return of -0.13 percent. The median return for the quarter for all 50 ETFs was 2.69 percent. In the first two weeks of the second quarter, the average return for the topfive ETFs was 2.79 percent; it had climbed to 7.22 percent by the end of the quarter. QQQQ gained 0.57 percent for the quarter vs. being up 0.19 percent at the beginning of the quarter. SPYadded an additional 1 percent from the second Friday to the end of the quarter. Figure 6 compares the initial and end-of-month returns for this quarter. Although Q4 2004 reinforced the tendency highlighted in “Following the money” for fourth quarters to have the best returns, it is worth noting that the period between the second Friday and the end of the quarter can be a volatile period. Figure 7 shows the relative price action based on percentage change for the top five ETFs on Oct. 8, 2004 from the start of the quarter to the end. Once October had passed the ETFs started climbing — except energy (XLE), which was flat.

What the performance says
First, the trading-range behavior of the broad market during the review period is evident. There was no follow-through movement (up or down) for two entire quarters. Nonetheless, the top-five ETFs had better-than-average returns (compared to QQQQ and SPY) from the second Friday to the end of the quarter four out of the six quarters. For swing traders, this suggests the top-five ETFs could offer better trading opportunities than QQQQ and SPY. When the market peaked in the first quarter of 2004, the initial top-two ETFs led the way down for the quarter (SMH
www.activetradermag.com • December 2005 • ACTIVE TRADER

Source: eSignal

68

FIGURE 6 BEGINNING VS. END OF SECOND QUARTER 2005 Only 44 percent of ETFs showed positive returns by the second Friday, but the early leaders performed well, as did ETFs overall. was 50th and EWZ was 44th). The same thing happened in the second quarter of 2004, except the HHH was in its own world while the ETFs ranked 2, 3, and 4 (EWY, SMH, and EWT) came in at or near the bottom for the quarter. On the other hand, in the third quarter of 2004, four of the initial top five landed in the top 10 at the end of the quarter. The early leaders can tell us a great deal about the quarter: If the early leaders begin to roll over, then the overall market can be in trouble. If the early leaders can maintain their relative performance, then the outlook is more positive. Considering how much the oil market has dominated the markets and looking over the big winners for the last six quarSource: eSignal ters, there should not be any surprise energy-related ETFs were leaders at the outset of three of the FIGURE 7 FOURTH QUARTER 2004 DAILY PERCENTAGE GAINS past six quarters. Similarly, healthcare and biotech ETFs This relative view of the top-five ETFs for the fourth quarter of 2004 shows that after (IBB, BBH, and XLV) made up October, the ETFs headed higher, with QQQQ coming in second. three of the initial top five in the second quarter of 2005 and returned an average of nearly 10 percent for the quarter. This research confirms what “Following the money” originally indicated: There is evidence that initial market trends can hold for an entire quarter. In addition, tracking the topfive leading ETFs can provide an indication of how the quarter may eventually unfold. Funds appear to move in and out of various groups based on a “theme” and reflect the herd mentality of market participants. This process of ranking ETFs at the start of the quarter is a simple but valuable analytical tool for staying on the right side of the market overall, as well as finding particular sectors and instruments with above-average trading opporSource: eSignal tunities.Ý
ACTIVE TRADER • December 2005 • www.activetradermag.com 69

TRADING Strategies

FOLLOW the leaders
Updating analysis of ETF quarterly performance reveals market patterns over the past year and a half.

BY THOM HARTLE

raders gravitate toward opportunity, and the easiest-to-exploit trading opportunities occur when the broad market is trending strongly and certain stock sectors or groups are attracting capital because of strong fundamentals. Regardless of the type of strategy you use, you can increase your chances of success by finding the best instruments to trade. Simple screening techniques can signal which groups are exhibiting the kind of relative strength that sets them apart from the broader market. The following analysis updates a screening process explained in two previous Active Trader articles — “Following the money: Quarterly momentum and ETFs” and “The ETF money trail” (see “Related reading”). “Following the Money: Quarterly Momentum and ETFs” details how comparing the closing prices of the 50 most-active exchange traded funds (ETFs) on the second Friday of a quarter
70 www.activetradermag.com • February 2006 • ACTIVE TRADER

T

TABLE 1 Q1 2004 to their closing prices at the end of The average return for the top-10 ETFs from Jan. 9, 2004 to March 31, 2004 was nearly identical to the S&P 500’s return. the previous quarter provides insights into potentially top-perName Symbol 1/9/04 3/31/04 Difference* forming ETFs, as well as the Merrill Lynch Semiconductor HOLDRS SMH 7.86% -4.80% -12.65% underlying strength of the market iShares MSCI Brazil Index Fund EWZ 7.65% -1.18% -8.82% as a whole. For example, the leading ETFs iShares MSCI Mexico Index Fund EWW 7.14% 21.55% 14.40% in the first two weeks of a quarter iShares MSCI Taiwan Index Fund EWT 6.32% 8.36% 2.05% could be a reflection of an investiShares MSCI Hong Kong Index Fund EWH 5.30% 6.00% 0.70% ment theme being pursued by Merrill Lynch Market Oil Service HOLDRS OIH 5.15% 13.27% 8.13% large money managers. The impliiShares MSCI Emerging Index Fund EEM 5.07% 7.06% 1.99% cation is that those leaders could Select Sector SPDR Fund — Technology XLK 5.00% -1.08% -6.08% dominate the entire quarter. Also, when a high percentage of iShares NASDAQ Biotechnology Index Fund IBB 4.46% 7.42% 2.96% the top-50 ETFs were up on the iShares Russell 2000 Growth Index Fund IWO 4.37% 5.47% 1.10% second Friday of the quarter, it was S&P 500 SPX 0.89% 1.29% 0.39% an indication there was wideAverage change for the top 10 0.38% spread demand for equities. On Percentage of top-50 ETFs that were bullish 84.00% the other hand, if only a small per*Results in the Difference column may vary due to rounding. centage of these ETFs were up on the second Friday, the likelihood of Source: data — eSignal a good quarter decreased. Comparing the performance of FIGURE 1 Q1 2004 RELATIVE PERFORMANCE the top-10 ETFs and the S&P 500 index (SPX) for each of the four 2004 quarters The bars at the bottom of the chart show the number of the top-10 ETFs that and the first two quarters of 2005 will were outperforming the S&P 500 at any given time. Seven of these ETFs had reveal whether these tendencies have outperformed the S&P 500 by the end of the quarter. Notice if a leader falls remained intact, or if the market is below the S&P 500, it is unlikely to recover. exhibiting other characteristics. The results have useful implications for both shorter-term trading and longer-term investors.

2004 quarterly performance
Table 1 lists the top-10 ETFs on Jan. 9, 2004 (the second Friday of the quarter) and their percentage returns as of the close on March 30, 2004 (the end of the quarter). The bottom of the table shows that on Jan. 9, 84 percent of the top-50 ETFs had positive returns, which indicated good demand at the start of the quarter and year. The best-performing ETF at the beginning of the quarter was the Merrill Lynch Semiconductor HOLDRS (SMH), which ended up making a big downswing, losSource: data — eSignal; chart — Excel ACTIVE TRADER • February 2006 • www.activetradermag.com 71

TABLE 2 Q2 2004 The S&P 500 was only marginally positive at the end of the second quarter, and only three of the top-10 ETFs outpeformed the index. Name Merrill Lynch Internet HOLDRS iShares MSCI South Korea Index Fund Merrill Lynch Semiconductor HOLDRS iShares MSCI Taiwan Index Fund NASDAQ-100 Index Tracking Stock Merrill Lynch Biotech HOLDRS Select Sector SPDR Fund — Technology iShares MSCI Emerging Index Fund iShares NASDAQ Biotechnology Index Fund S&P 500 Average change for the top 10 Percentage of 50 bullish
*Results in the Difference column may vary due to rounding.

Symbol 4/8/04 6/30/04 Difference* HHH EWY SMH EWT QQQQ BBH XLK EEM IBB SPX 9.57% 5.29% 3.92% 3.28% 3.07% 3.07% 2.93% 2.82% 2.74% 2.48% 1.16% 76.00% 26.48% -13.66% -4.23% -11.17% 5.30% 3.06% 8.52% -7.90% -2.83% 0.58% 1.30% 16.92% -18.94% -8.15% -14.45% 2.23% -0.01% 5.59% -10.72% -5.58% -1.90% 0.13% -3.50% ing 12.65 percent from Jan. 9 to March 30. Overall, however, seven of the top-10 ETFs followed through with gains for the remainder of the quarter that were greater than those of the S&P 500. Figure 1 shows the percentage changes for the top-10 ETFs (plus the S&P 500 index) from Table 1. The bars at the bottom of the chart indicate how many of the 10 ETFs are outperforming the S&P 500. There was a tendency for an ETF whose relative performance dropped below the performance of the S&P 500 (for example, SMH, EWZ, and XLK) to continue to lag the index. Also, notice that as the market began to falter in early March the entire group declined, but the final leader (EWW, the iShares MSCI Mexico Index Fund) diverged from the group and ended up making a new high for the quarter. Table 2 shows the top-10 performing ETFs at the start of 2004’s second quarter and their respective outcomes. The top early leaders did not fare as well this quarter. The average return from April 8 to June 30 was a 3.5-percent loss, while the S&P 500 was essentially unchanged (a .13-percent gain). This goes against past tendencies, in that the early leaders floundered even though 76 percent of the top-50 ETFs had positive returns on April 8 — a bullish scenario. The lone standout was the Merrill Lynch Internet HOLDRS (HHH), which tacked on an additional 16.92 percent for the quarter. Figure 2 shows how this quarter progressed. Before the first month ended, six of the 10 ETF’s percentage performance lines were below the S&P’s, and when the index went negative for the

iShares Goldman Sachs Software Index Fund IGV

Source: data — eSignal

FIGURE 2 Q2 2004 RELATIVE PERFORMANCE One month into the quarter only three of the top-10 ETFs were outperforming the S&P 500 — not a good sign for the market as a whole. Generally, when five or more of the early leaders have negative returns in the first month of the quarter, the outlook for the remainder of the quarter is poor.

Source: data — eSignal; chart — Excel

72

www.activetradermag.com • February 2006 • ACTIVE TRADER

TABLE 3 Q3 2004 The average performance of the top-10 ETFs was far superior to the S&P 500’s performance during this quarter. Name iShares MSCI Malaysia Index Fund Select Sector SPDR Fund — Energy Select Sector iShares Dow Jones U.S. Real Estate Index Fund Symbol 7/9/04 9/30/04 Difference* EWM XLE 4.96% 1.43% 4.19% 10.96% -0.78% 9.53%

IYR 1.31% 6.86% 5.55% quarter, the losses mounted for the iShares MSCI Hong Kong Index Fund EWH 1.18% 9.77% 8.59% bottom three ETFs. As in Figure 1, iShares Lehman this chart shows leaders that fall 20-Year Treasury Bond Fund TLT 0.87% 5.50% 4.63% out of favor are unlikely to recover. Also, when five or more early iShares Lehman 7-10 Year Treasury Bond Fund IEF 0.76% 3.41% 2.65% leaders have negative returns in the quarter’s first month, the outiShares MSCI Brazil Index Fund EWZ 0.75% 26.04% 25.29% look for the quarter is poor. Merrill Lynch Market Oil Service HOLDRS OIH 0.43% 13.11% 12.68% Table 3 shows 2004’s third-quariShares Lehman ter performance. At the start of the 1-3 Year Treasury Bond Fund SHY 0.13% 0.40% 0.27% quarter, only 18 percent of the topSelect Sector SPDR Fund — Utilities XLU -0.72% 5.43% 6.14% 50 ETFs had gains (and were, in S&P 500 SPX -2.46% -2.36% 0.10% fact, the first nine ETFs in the topAverage change for the top 10 7.46% 10 list). This is an indication the quarter is off to a poor start. Percentage of 50 bullish 18.00% However, by the end of the quar*Results in the Difference column may vary due to rounding. ter, all 10 ETFs had gains, and only one (EWM) had lost ground from Source: data — eSignal July 9. The average return for the top 10 was 7.46 percent. (Also, FIGURE 3 Q3 2004 RELATIVE PERFORMANCE notice three of Table 3’s top-performing ETFs were fixed-income funds. While There were only three days during this quarter when all 10 ETFs were underperforming the S&P 500. they are not equity oriented, they have defensive characteristics that added to the bottom line in this case.) Figure 3 shows how the early leaders outperformed the S&P 500, despite some of them having gone negative (dropping below 1) midway through the quarter before climbing back into the black. With the exception of three days in August, all the ETFs outperformed the S&P 500 for the entire period. The 2004 fourth-quarter results are displayed in Table 4 (p. 46). On the second Friday of the quarter, 78 percent of the 50 most-active ETFs were up, which was a sign of optimism for the end of the year. There is a strong seasonal tendency for the fourth quarter to do well after October has passed. Figure 4 shows how troublesome October was
Source: data — eSignal; chart — Excel

ACTIVE TRADER • February 2006 • www.activetradermag.com

73

TABLE 4 Q4 2004 After a bullish start, both the top-10 ETFs and the S&P 500 posted strong returns this quarter, with the ETFs closing the quarter slightly stronger. Name iShares MSCI Brazil Index Fund iShares MSCI South Korea Index Fund Select Sector SPDR Fund — Energy Select Sector iShares Goldman Sachs Software Index Fund iShares MSCI Taiwan Index Fund Merrill Lynch Market Oil Service HOLDRS Merrill Lynch Telecom HOLDRS iShares MSCI Germany Index Fund iShares MSCI Japan Index Fund iShares MSCI Emerging Index Fund S&P 500 Average change for the top 10 Percentage of 50 bullish
*Results in the Difference column may vary due to rounding.

Symbol 10/8/04 12/31/04 Difference* EWZ EWY XLE IGV EWT OIH TTH EWG EWJ EEM SPX 4.64% 4.60% 3.61% 3.40% 3.36% 3.16% 3.10% 2.79% 2.68% 2.43% 0.68% 78.00% 19.96% 17.00% 3.98% 23.05% 12.71% 4.12% 1.57% 18.14% 12.58% 17.01% 8.67% 15.32% 12.40% 0.37% 19.65% 9.35% 0.97% -1.53% 15.35% 9.90% 14.58% 8.00% 9.63% for the market as a whole. However, when it was over funds flowed into the market. At the end of October, only four ETFs were outperforming the S&P 500, which is not bullish. But considering how October tends to be a bad month for the stock market and the rest of the quarter tends to be strong, an “allclear” signal was given when the S&P 500 and eight of the 10 ETFs moved into positive territory (climbing above 1) in early November. Figure 4 shows again that when an ETF begins to lag the S&P 500 it tends not to outperform it by the end of the quarter. Nonetheless, by the end of the quarter seven of the 10 had outperformed the S&P 500 and only one (TTH) lost money from Oct. 8. Looking at 2004 as a whole, the market swung up and down in a broad trading range. With the exception of the fourth quarter, a high percentage of ETFs with positive return at the start of a quarter did not necessarily generate follow-through upside action for the remainder of the quarter. On the other hand, the weak start for the third quarter (only 18 percent of the 50 most-active ETFs had positive returns) was reversed by the top leaders.

Source: data — eSignal

FIGURE 4 Q4 2004 RELATIVE PERFORMANCE Despite a volatile October, seven of the top-10 ETFs outperformed the S&P 500 this quarter, and only one (TTH) posted a negative return.

2005 performance
Table 5 shows the first quarter for 2005. At the start, only 14 percent of the 50 most-active ETFs were positive on the second Friday of January 2005. Those seven are in the top-10 list, and two of them are fixed-income instruments. The group’s returns were wide-ranging over the quarter. The top performer, Select
www.activetradermag.com • February 2006 • ACTIVE TRADER

Source: data — eSignal; chart — Excel

74

TABLE 5 Q1 2005 Reviewing the best-performing ETFs from Jan. 14, 2005 shows that despite the market’s weak start, the average end-of-quarter return for the top-10 ETFs was nearly 2 percent better than the S&P 500’s return. Name iShares MSCI Malaysia Index Fund iShares MSCI South Korea Index Fund Sector SPDR Fund-Energy Select Sector (XLE), gained an additional 17.65 percent from Jan. 14, while the poorest performer, the iShares MSCI Malaysia Index Fund (EWM), reversed and lost 9.79 percent from Jan. 14. Figure 5 shows two of the top three performers were energy oriented — XLE and the Merrill Lynch Oil Service HOLDRS (OIH). Most of the ETFs peaked in early March and trended lower for the rest of the quarter. The S&P 500 led the way down — the bar count at the bottom of the chart began rising and reached 10 in mid-March, indicating all 10 ETFs were outperforming the S&P 500 during the decline. Table 6 (p. 48) shows only 44 percent of the top-50 ETFs were bullish at the start of the second quarter 2005. The S&P 500 was flat for the quarter, while the 10 ETFs gained an average of 4.13 percent from April 8 to the end of the quarter. Biotech was the big winner: The Merrill Lynch Biotech HOLDRS (BBH) added an additional 16.23 percent from April 8. Figure 6 shows the majority of the 10 ETFs did better than the S&P 500 for the entire quarter — the histogram count never dropped below eight. (The Merrill Lynch Biotech HOLDRS [BBH] was off in its own world.) iShares Lehman 20 Year Treasury Bond Fund Merrill Lynch Market Oil Service HOLDRS Select Sector SPDR Fund — Consumer Staples iShares Lehman 7-10 Year Treasury Bond Fund Select Sector SPDR Fund — Energy Select Sector iShares Lehman 1-3 Year Treasury Bond Fund Merrill Lynch Retail HOLDRS iShares MSCI Japan Index Fund S&P 500 Average change for the top 10 Percentage of 50 Bullish
Source: data — eSignal

Symbol 1/14/05 3/31/05 Difference EWM EWY TLT OIH XLP IEF XLE SHY RTH EWJ SPX 4.76% 2.05% 1.83% 1.70% 0.48% 0.44% 0.39% -0.07% -0.27% -0.64% -2.26% 14.00% -5.03% 8.75% 0.88% 13.02% -0.26% -1.48% 18.03% -0.69% -2.85% -3.94% -2.59% -9.79% 6.70% -0.95% 11.32% -0.74% -1.92% 17.65% -0.61% -2.58% -3.30% -0.33% 1.58%

FIGURE 5 Q1 2005 RELATIVE PERFORMANCE The top three ETFs this quarter were XLE, OIH, and EWY, the first two of which are energy oriented. Seven of the 10 ETFs outperformed the S&P 500, which posted a slight loss.

Market characteristics and trading implications
The average return for the top-10 ETFs from the second Friday of the quarter to the end of the quarter was better than the S&P 500 return in four of the six quarters. (The S&P 500 bested the ETFs

Source: data — eSignal; chart — Excel

ACTIVE TRADER • February 2006 • www.activetradermag.com

75

TABLE 6 Q2 2005 The S&P 500 gained less than 1 percent from April 8 to June 30, while the average gain for the top-10 ETFs came in at 4.13 percent.

by a mere 0.01 percent in Q1 2004, and in Q2 2004 the ETFs posted an Name Symbol 4/8/05 6/30/05 Difference* average loss of 3.50 percent while iShares MSCI South Korea Index Fund EWY 3.74% 0.60% -3.14% the S&P was up 0.13 percent.) iShares NASDAQ Biotechnology Index Fund IBB 2.86% 6.85% 3.98% For the entire six-quarter period, the sum of the averages for the topiShares MSCI Emerging Index Fund EEM 2.71% 5.92% 3.21% 10 ETFs was 19.67 percent, while the Merrill Lynch Biotech HOLDRS BBH 2.60% 18.83% 16.23% sum for the S&P 500 was 9.15 perSelect Sector SPDR Fund — Health Care XLV 2.01% 3.92% 1.91% cent. Execution costs, slippage, and Merrill Lynch Pharmaceutical HOLDRS PPH 1.75% 1.92% 0.17% dividends were not considered in iShares MSCI Hong Kong Index Fund EWH 1.30% 7.91% 6.60% this review. Select Sector SPDR Fund — Utilities XLU 1.03% 8.27% 7.24% Tracking the top 10 performers in the first two weeks of a quarter proiShares MSCI EAFE Index Fund EFA 1.03% -1.07% -2.10% vides a useful watch list for swing Merrill Lynch Utilities HOLDRS UTH 0.89% 8.09% 7.20% trading. For example, if an ETF’s S&P 500 SPX 0.05% 0.91% 0.86% performance starts to lag the S&P Average change for the top 10 4.13% 500, you can assume this ETF has Percentage of 50 bullish 44.00% lost its luster. Such ETFs are likely *Results in the Difference column may vary due to rounding. weaker candidates for long-trade setups, but may offer potential on Source: data — eSignal the short side (especially if the overall market is weak). Also, for the past six quarters, either energy or biotech (sometimes both at the FIGURE 6 Q2 2005 RELATIVE PERFORMANCE same time) have been in the group of The S&P ultimately finished the quarter with a minor gain. However, eight of top-10 ETFs. This simple screening the top-10 ETFs came in with stronger performance, led by the Merrill Lynch process would have alerted you to their Biotech HOLDRS (BBH). strength early on.

Related reading
Other ETF articles by Thom Hartle:
“Following the money: Quarterly momentum and ETFs” Active Trader, June 2004. “The ETF money trail” Active Trader, December 2005. You can purchase and download past Active Trader articles at www.activetradermag.com/purchase_articles.htm.
Source: data — eSignal; chart — Excel

76

www.activetradermag.com • February 2006 • ACTIVE TRADER

TRADING Strategies

The intraday MFI
The bar-to-bar comparisons of volatility and volume in the Market Facilitation Index make it possible to logically determine when the indicator is relatively high and price follow-through is likely.

BY THOM HARTLE

T

here are different ways traders attempt to identify “quiet periods” that have the potential to be followed by bursts of price activity. Some people look for narrow consolidations, for

example. The Market Facilitation Index (MFI) is an indicator that can serve the same role — mathematically. In “The Market Facilitation Index” (Active Trader, October 2004), the MFI was applied to 45-minute bars of the E-

FIGURE 1 MFI: TRACKING VOLUME AND VOLATILITY High Market Facilitation Index readings indicate low-volume and sideways movement. In this case, the MFI peaked right before a rally in the E-Mini S&P 500 futures.

Mini S&P futures (ES) as a way to determine periods of intraday price action from which short-term trends might emerge. The following analysis updates that study of the E-Mini S&P and also applies the same analysis to the E-Mini Nasdaq 100 (NQ) and E-Mini Russell 2000 (ER2) futures contracts to look for unique opportunities in these three markets.

MFI background
The MFI was introduced by Bill Williams, Ph.D., in his book Trading Chaos (John Wiley & Sons, 2004). The MFI is a bar’s price range divided by its volume: MFI = (High - low)/Volume Low MFI readings occur when volume is high relative to the range. If volume subsides and the bar ranges remain relatively constant, the MFI reading will climb. Similarly, the MFI reading will climb when the ranges increase but volume does not. This basic calculation leads to very small values: If a bar’s range in the EMini S&P 500 was 4.25 points and the volume was 90,782 contracts, the MFI would be 4.25/90,782 = 0.0000468154. Accordingly, the raw MFI value is multiplied by 1 million to make the number easier to use (in this case, resulting in an MFI reading of 46.82). Figure 1 shows one day’s activity in
www.activetradermag.com • September 2006 • ACTIVE TRADER

Source: CQGNet, Inc.

77

FIGURE 2 RANGE VS. VOLUME 45-minute bars of the E-Mini S&P. At bar A, the market trended down from the previous day’s final 45-minute bar, the volume was 128,402 contracts, and the MFI was 21.42. With bar B, the market edged lower, the bar’s range was slightly wider, and the volume decreased to 115,065; the MFI climbed to 30.42. Bar C was an inside bar (a lower high and a higher low than the preceding bar) and the volume dropped further to 62,272 contracts; the MFI rose to 36.13. Bar D established another new low for the day. It had a larger range than bar C but the volume dropped to 55,162 and the MFI jumped to 49.85. Bar E, which began at 11:30 CT, traded above bar D. It had a small range and its volume was the lowest of the day (21,711 contracts); its MFI reading was 80.60. Bar F, which began at 12:15 CT, had a larger range than bar E and volume rose to 26,487; it had the highest MFI reading of the day — 84.95. Note that bars E and F encompass lunchtime in Chicago, where the E-Mini futures are traded. From this point, the market rallied. Bar G kicked off a sharp rally. It had a six-point range and volume of 136,090 — the highest of the day. The MFI dropped to 44.09. Bar H marked another new high for the day, but the range narrowed and the volume was still high at 88,768 contracts traded; the MFI dropped to 33.80. Finally, bar I was a narrow-range bar with higher volume (109,751 contracts) than bar H, and the MFI dropped to 20.50. High MFI readings indicate low-volume, narrow-range trading periods. Markets tend to move from low-volume, sideways activity into trends accompanied by heavy volume, and then back again into trading ranges. During this period the three-month average monthly volume (immediately below price) increased while the three-month average monthly range (bottom) fell.

Source: CQGNet, Inc.

Low MFI readings indicate high volume, which means traders have taken their positions in the market; there is less likely to be follow-through once traders have placed their trades. In the Figure 1 example, the greatest opportunity to capture a good intraday trend followed the higher MFI readings, not the low ones. Now let’s analyze three markets to see exactly what happens after high MFI readings.

Analyzing high MFI readings
The original study from the October 2004 article analyzed 45-minute bars of the E-Mini S&P futures from Jan. 2, 2004 to June 10, 2004. The new study covers 125 trading days from Oct. 3, 2005 to March 31, 2006. Both studies use pit-trading hours, which open at 8:30 a.m. CT and close at 03:15 p.m. CT. The mean (average) MFI

TABLE 1 THE MFI BEFORE THE EMPLOYMENT REPORT The day before the employment report (March 9, 2006), the MFI readings were unusually high. Time (CT) MFI 8:30 83.55 9:15 155.52 10:00 115.44 10:45 236.79 11:30 242.84 12:15 230.65 13:00 382.70 13:45 211.80 14:30 163.37

ACTIVE TRADER • September 2006 • www.activetradermag.com

78

FIGURE 3 PRE-EMPLOYMENT REPORT ACTION On March 9, 2006, the market trended lower but volume was very low. High MFI readings were the result. minute bar MFI readings was used to determine what constitutes a “high” MFI reading. In this case, any MFI reading greater than the mean MFI plus one standard deviation was considered high. However, we need to look a little more closely at the data before finalizing this threshold. The standard deviation for the current study was 24.91, compared to 23.98 for the previous study. The fact that the standard deviation increased slightly while the mean MFI reading dropped 21 percent implies some outliers (extreme, anomalous values) are skewing the MFI readings higher. Sorting the MFI readings from high to low revealed that March 9, 2006 — the day before the monthly employment report release — produced some very high MFI readings relative to the rest of the analysis period. Table 1 lists the MFI readings and Figure 3 is a chart of that day. Given that days immediately preceding important economic reports are often exceptionally quiet, the high MFI readings are not surprising. Removing the March 9 data lowers the mean from 49.75 to 48.51, but the standard deviation drops from 24.91 to 19.48. Adding the new standard deviation (19.48) to the new mean (48.51) results in a high MFI threshold of 67.99. Table 2 shows the number of times the MFI exceeded 67.99 during the new analysis period. Table 3 shows comparable statistics from the original study,

Source: CQGNet, Inc.

from Jan. 2, 2004 to June 10, 2004 was 63.16. For the new period, the mean MFI dropped to 49.75 — a 21-percent decline. Figure 2 offers an explanation as to why this change occurred. It’s a monthly chart of the E-Mini S&P 500 from January 2004 to March 2006 with volume and monthly range (and their TABLE 2 HIGH MFI READINGS

respective three-month simple moving averages). It shows the average monthly volume is higher and the average monthly range is lower in the last six months compared to the first six months. Smaller ranges and higher volume translate into lower MFI readings. The standard deviation of the 45-

Tallying the number of times the MFI exceeded 67.99 for each 45-minute bar between Oct. 3, 2005 and March 31, 2006 shows the high readings tended to cluster mostly between 10:45 and 13:00. Time (CT) MFI> 67.99 8:30 0 9:15 2 10:00 5 10:45 25 11:30 46 12:15 38 13:00 24 13:45 16 14:30 0

TABLE 3 PREVIOUS STUDY This summary of high MFI readings from Jan. 2, 2004 to June 10, 2004 is from the original study published in October 2004. The readings were similar to those in Table 2, except there were fewer high readings for the 13:45 bar. Time (CT) MFI> 87 8:30 0 9:15 0 10:00 6 10:45 27 11:30 47 12:15 54 13:00 20 13:45 4 14:30 0

79

www.activetradermag.com • September 2006 • ACTIVE TRADER

FIGURE 4 STUDY COMPARISON when 87 was the MFI threshold. One thing the two studies have in common is the MFI readings for the first and last 45 minutes of the trading day never exceeded the MFI high threshold. This makes sense, as the beginning and end of the trading day have the largest volume — participants react to overnight news in the morning and adjust to the day’s trading going into the close. One difference between the studies is the greatest number of high MFI readings occurred at the 11:30 bar in the new study and the 12:15 bar in the original study. Also, the new study suggests there are more high MFI readings during the afternoon (specifically, the 13:00 and 13:45 bars) compared to 2004. Figure 4 compares the two studies side by side. The chart implies ranges are not expanding and volume is not climbing in the latter part of the day. In other words, the opportunities implied by a high MFI reading at one bar are not developing after a series of consecutive high MFI readings. The original study required the MFI to exceed 87, while the new study required the indicator to surpass 67.99. The 13:00 and 13:45 bars had higher MFI readings during the second study period.

FIGURE 5 MFE-MAE STUDY COMPARISON The 12:15 bar has a higher average MFE for the 2005-2006 study (purple) and a more negative MAE (burgundy) than the same bar in the 2004 study (gold for the MFE and light blue for the MAE).

Measuring price reaction to high MFI readings
Now it’s time to look at the followthrough price movement once a high MFI reading has occurred. To do this, the maximum favorable excursion (MFE) and maximum adverse excursion (MAE) are calculated after each bar with a high MFI reading, as follows: Calculating the MFE: 1. If the MFI at the close of a bar is greater than 67.99, and if the bar closes above its midpoint (implying bullish momentum), the difference between the close and the next bar’s high is calculated. 2. If the bar closes at its mid-point or lower (implying bearish momentum), the difference between the close and the next bar’s low is measured. The idea is the market may be leaning in the direction of an emerging trend following the high MFI reading.

Calculating the MAE: 1. If the signal is for an up move (a close above the midpoint), the difference between the close and the next bar’s low is calculated. 2. If the signal is for a down move (a close below the midpoint), the difference between the close and the next bar’s high is calculated. Table 4 shows the median MFE and

MAE values for each 45-minute bar in the two studies. Figure 5 displays the average MFE and MAE for each study. Figure 6 shows the respective net differences between MFEs and MAEs for the two studies.

E-Mini S&P results
Because the MFI exceeded 67.99 at the close of the 9:15 bar only two times in six

ACTIVE TRADER • September 2006 • www.activetradermag.com

80

TABLE 4 MFE AND MAE STATISTICS Comparing the median maximum favorable excursion (MFE) and maximum adverse excursion (MAE) figures for each 45minute bar in the two studies indicates the largest moves occur around lunchtime in Chicago. New study: 10/03/2005 to 3/31/2006 Occurrences Time MFE MAE 2 9:15 0.3750 -1.6250 5 10:00 1.5000 -0.6500 25 10:45 1.4200 -1.3400 46 11:30 1.6630 -1.3696 38 12:15 2.2434 -2.0461 24 13:00 1.7292 -1.3542 16 13:45 1.5156 -1.8125 Original study: 1/02/2004 to 6/10/2004 Occurrences Time MFE MAE 0 9:15 0.0000 0.0000 6 10:00 3.1250 -0.8750 27 10:45 1.5000 -0.5000 47 11:30 0.7500 -2.0000 54 12:15 1.7500 -1.1250 20 13:00 1.2500 -1.5000 4 13:45 1.3750 -0.7500

TABLE 5 E-MINI NASDAQ 100 MFI READINGS The number of occurrences is highest at lunchtime in Chicago. Time (CT) MFI > 38.44 8:30 1 9:15 4 10:00 7 10:45 18 11:30 39 12:15 39 13:00 22 13:45 13 14:30 1

direction after this bar. The current study produced an MFE of 1.42 and an MAE of The average net points for the MFE of the -1.34, so this tendency e-mini Nasdaq 100 futures contract is higher seems to have dissipated. than the E-Mini S&P 500 contract. In the current study the MFE is largest after the Occurrences Time MFE MAE 12:15 bar, coming in at 1 8:30 3.000 -1.000 2.2434 — but at -2.0461, the 4 9:15 4.000 -2.125 MAE after this bar also is 7 10:00 3.143 -1.357 the largest. This indicates an 18 10:45 3.083 -1.917 increase in volatility but not 39 11:30 3.090 -2.846 necessarily directional 39 12:15 3.038 -3.859 movement. (Nonetheless, 22 13:00 3.500 -2.500 the edge is towards the 13 13:45 2.231 -3.346 direction of the close rela1 14:30 0 0 tive to the bar.) The final two 45-minute bars also had higher MFEs in the new study, but the MAE for the months in the second study and zero times in the first study, this bar can be final bar in the new study was ignored. Although high MFI readings for -1.8125 points compared to just -0.75 the 10:00 bar also were rare (only five points in the original study. More imporand six occurrences in the new and old tantly, the MFE for the final bar was studies, respectively), the second study 1.5156 for the new study. In other words, showed a drop in both the MFE and the setup saw a negative return (+1.5156 MAE for this bar. The 10:45 bar has near- vs. -1.8125). The fact that over the 124 days of the ly the same MFE readings in both studies, but the MAE reading nearly tripled current study (March 9, 2006 was in size in the new study, implying non- removed) there were 133 high MFI readdirectional price movement following ings from the 10:45 through 13:00 bars — all of which have higher MFEs than the high MFI reading. The 11:30 bar had an MFE of 0.75 and MAEs — indicates the high MFI reading MAE of -2.00 in the first study, which could be the basis of a trading strategy. indicated the market tended to reverse In other words, the higher MFEs suggest
81

TABLE 6 E-MINI NASDAQ 100 MFE AND MAE READINGS

more follow-through price movement is occurring in the direction of the bar’s close (above the midpoint indicating an up move and below the midpoint indicating a down move) after these bars.

E-Mini Nasdaq 100 and Russell 2000 results
The same analysis was applied to E-Mini Nasdaq 100 (NQ) and E-Mini Russell 2000 (ER2) futures. The E-Mini Nasdaq 100 futures contract had a median MFI reading of 27.37 and a standard deviation of 11.07, so an MFI reading greater than 38.44 was considered high. Table 5 details the results for October 2005 through March 2006 (again, with March 9 removed). The MFI did exceed 38.44 once each in the first and last 45 minutes of the day. The greatest number of high MFI readings occurred in the 11:30 and 12:15 bars (both had 39). Table 6 details the MFE and MAE numbers. The peak MFE (4.00 points) occurred after the the 9:15 bar, but there were only four occurrences. After the 12:15 bar the MAE was larger (-3.859 points) than the MFE (3.038 points). Even more challenging for directional traders is the price behavior after the 13:45 bar — an MAE of -3.346 vs. an MFE of 2.231, indicating a reversal tendency during the afternoon trading.

www.activetradermag.com • September 2006 • ACTIVE TRADER

FIGURE 6 NET DIFFERENCE The 11:30 bar in the latest study has a larger MFE and MAE than the original study. The 12:15 bar has a larger MFE and a smaller MAE. The 13:00 bar has a larger MFE and MAE.

E-Mini Russell 2000 results
Tables 7 and 8 contain the analysis of the E-Mini Russell 2000. The median MFI reading was 26.13 and the standard deviation was 8.89, making the MFI high threshold 35.02. Table 7 summarizes the results for October 2005 through March 2006, with trading on March 9 removed. Of the three markets, the E-Mini Russell had the greatest number of high MFI readings in the first and last 45minute bars. And as was the case with the other two markets, the peak number of high MFI readings occurred around lunchtime in Chicago. TABLE 7 E-MINI RUSSELL 2000 AND THE MFI In the E-Mini Russell 2000 futures, the MFI exceeded the high reading in the first 45 minutes of trading four times. Time (CT) MFI > 35.02 8:30 4 9:15 1 10:00 8 10:45 25 11:30 43 12:15 40 13:00 14 13:45 11 14:30 8

From the close of the 10:00 bar to the close of the 12:15 bar, the MFEs edge out the MAEs, with the 10:45 bar having the widest margin. The afternoon trading in the E-Mini Russell indicates a tendency towards reversals, as the MAEs are larger than the two MFEs.

The analysis was built upon a very simple signal. Additional rules could be included to fine-tune the identification of a reliable trend following a high MFI reading. An additional filter could be applied that might shift the high MAE readings over to the MFE column.

Related reading
“The squat bar” by Thom Hartle (Active Trader, January 2005). Traders often anticipate that trend moves will follow low-volume, lowvolatility periods. The squat-bar pattern shows high-volume bars can also lead to tradable price moves. The MFI is used as one of two parameters to classify price bars into four categories. “The Market Facilitation Index” By Thom Hartle (Active Trader, October 2004). Traders are always looking for ways to determine when the market is poised to make a move. The Market Facilitation Index is a tool you can use to identify market lulls that can precede price trends. You can purchase and download past articles at www.activetradermag.com/purchase_articles.htm.

The lessons of the data
The analysis provides some useful insights into the ways these markets typically move intraday. Not surprisingly, all three tend to slow down during lunchtime in Chicago. However, the S&P 500 tended to trend more during this period in the new study compared to the original 2004 study. The E-Mini Nasdaq 100 futures had a higher MAE than MFE following the close of the 12:15 bar. All three markets had larger MAEs than MFEs following the close of the 13:45 bar. TABLE 8 E-MINI RUSSELL 2000 MFE AND MAE READINGS During the afternoon, the MAE tended to be worse than the MFE for the E-Mini Russell 2000 futures. Occurrences 4 1 8 25 43 40 14 11 8 Time 8:30 9:15 10:00 10:45 11:30 12:15 13:00 13:45 14:30 MFE 1.425 0 1.238 1.344 1.535 1.638 0.893 1.236 0 MAE -1.250 -1.700 -0.825 -0.952 -1.349 -1.600 -1.757 -1.355 0

ACTIVE TRADER • September 2006 • www.activetradermag.com

82

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->