Professional Documents
Culture Documents
Speed On The Sand
Speed On The Sand
All-Weather racing in Britain has always been looked upon as the poor cousin of turf racing, mainly due to the low quality of racehorse it attracts and the dismal prize-money which is attributed to the majority of the races under this particular code. Relations are improving, thanks to the emergence of the universally acclaimed Polytrack surface, which is generally accepted to be a safer and fairer surface than the older sand-based compositions Equitrak and Fibresand. Prizemoney is considerably higher too, which means a better class of runner is attracted which, in turn, leads to more interest from punters and the media, and this has helped raise the profile of racing on artificial surfaces in this country. From experience of compiling my own speed ratings for many years, I know that the art of rating performances by the use of the clock can reap rich rewards, and the one area of speed ratings which stands out from the rest has always been All-Weather racing. There seems to be fewer races which develop into tactical affairs on sand than on turf, possibly due to the general lower class of race on the all-weather courses, with many races of these events being run at a true gallop right from the off - this results in more races being available for time analysis which leads to more reliable and accurate speed figures. So, if speed ratings are to be presumed more accurate on sand than on turf, this should be reflected in the results of Topspeed, the Racing Posts very own speed rating compiler. If we run through the last 18 years of flat season data using the RSB program, searching for the results of all horses which were toprated by Topspeed, and had an SP of shorter than 4/1 (this is to concentrate on only those horses which were deemed good enough on form to go close), we arrive at the following results.
SURFACE WINS RUNS STRIKE% LSP LSP%
All Turf AW
Not exactly the result I was expecting, with the turf results showing a better figure on both of the important factors, the strike rate and LSP%. This seemed to fly in the face of logic, so I delved deeper to try and find out why the results had been the opposite of what I thought they would be.
I decided to take a closer look at the all-weather results, and the table below shows the results from each of the last 16 seasons.
YEAR WINS RUNS STRIKE% LSP LSP%
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
30 147 168 183 247 236 281 303 296 392 448 538 424 327 354 541
40.00 38.10 38.69 36.07 38.87 30.08 36.65 32.67 37.50 33.67 38.62 36.80 35.61 35.17 34.18 29.02
4.19 8.46 1.31 0.44 6.66 -12.57 4.83 -35.91 18.61 -20.10 7.68 4.88 8.26 19.35 2.00 -96.61
13.97 5.76 0.78 0.24 2.70 -5.33 1.72 -11.85 6.29 -5.13 1.71 0.91 1.95 5.92 0.56 -17.86
TOTAL
1726
4915
35.12
-78.52
-1.60
For some reason, 2004 resulted in a heavy loss not only did the strike rate drop from a supremely consistent mid-30s figure to a little over 29%, but the LSP loss was almost 100 points. In fact, were it not for this particular season, the total LSP would have been in profit rather than loss. Why did the results in 2004 suddenly take such a dramatic downturn? My immediate theory was targeted towards the arrival of the Polytrack surface at Wolverhampton racecourse, which replaced the original Fibresand. The old surface had become dangerously in need of replacement, which resulted in some unreliable results, and when the new track was laid later in the year the early results were once again untrustworthy due to Polytrack being a surface which encourages falsely run events. Therefore its not surprising to find that the results from Wolverhampton were largely responsible for the poor results from 2004. The three all-weather tracks are analysed individually below.
WOLVERHAMPTON
YEAR WINS RUNS STRIKE% LSP LSP%
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
0 29 41 39 31 39 56 75 66 42 45 50
0.00 30.85 39.42 39.39 33.70 30.95 35.67 35.21 40.99 32.81 35.16 24.75
-1.00 -4.53 15.49 7.58 2.35 -19.05 -12.90 -11.78 23.85 -2.43 -5.14 -51.43
-100.00 -4.82 14.89 7.66 2.55 -15.12 -8.22 -5.53 14.81 -1.90 -4.02 -25.46
TOTAL
513
1505
34.09
-58.99
-3.92
LINGFIELD
YEAR WINS RUNS STRIKE% LSP LSP%
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
5 19 32 33 35 17 41 32 44 55 68 60 32 33 33 43
15 52 85 91 101 78 101 115 108 145 161 166 95 101 112 170
33.33 36.54 37.65 36.26 34.65 21.79 40.59 27.83 40.74 37.93 42.24 36.14 33.68 32.67 29.46 25.29
-0.82 3.23 7.68 -3.76 -11.48 -21.75 10.01 -30.56 17.28 14.44 17.43 4.22 5.94 -1.87 -14.69 -48.36
-5.47 6.21 9.04 -4.13 -11.37 -27.88 9.91 -26.57 16.00 9.96 10.83 2.54 6.25 -1.85 -13.12 -28.45
TOTAL
582
1696
34.32
-53.06
-3.13
SOUTHWELL
YEAR WINS RUNS STRIKE% LSP LSP%
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
7 37 33 33 61 25 21 28 36 38 49 63 53 40 43 64
46.67 38.95 39.76 35.87 42.07 39.06 27.63 31.46 37.50 31.40 37.69 39.62 31.55 40.82 37.72 37.87
5.01 5.23 -6.37 4.20 19.14 13.71 -20.67 -12.93 -1.02 -15.49 3.15 12.44 -21.53 23.65 21.83 3.18
33.40 5.51 -7.67 4.57 13.20 21.42 -27.20 -14.53 -1.06 -12.80 2.42 7.82 -12.82 24.13 19.15 1.88
TOTAL
631
1714
36.81
33.53
1.96
Polytrack was introduced to Lingfield in November 2001 and to Wolverhampton in 2004. It seems perfectly logical to me to assume that the poor recent Topspeed results from these two tracks is down to this new speed-rating unfriendly surface. Southwell, after all, has continued to produce consistent figures in recent years, as has turf. So from now in this study, we will consider only those races which took part on non-Polytrack all-weather surfaces (Equitrack & Fibresand). Experience has taught me that race distance is very much a key factor in the study and use of speed ratings; the longer the race, the less likely that a true pace will ensue, leading to slow times and unreliable form. With this in mind, well now look at the stats from the last 5 seasons of these Topspeed selections on non-Polytrack all-weather surfaces (SP<=4/1) broken down by distance
DISTANCE
WINS
RUNS
STRIKE%
LSP
LSP%
<=10f 10f>
466 145
1267 441
36.78 32.88
65.04 -44.55
5.13 -10.10
TOTAL
611
1708
35.77
20.49
1.20
As can be seen, quite clearly a marked difference of fortunes over long and short distances. There is only 3.9% between them in terms of strike rate but the considerable gulf in LSP% shows that the runners over the longer trips are bad value. Concentrating on the runners which contested races over 10f or less, the yearly breakdown since 2000 looks like this.
YEAR
WINS
RUNS
STRIKE%
LSP
LSP%
159 112 59 74 62
TOTAL
466
1267
36.78
65.04
5.13
The main thing to notice here is the poor return from 2004, but the vast majority of these losses came from Wolverhampton in the months leading up to the laying of the new Polytrack surface. The Fibresand at Wolverhampton was in no fit state for horse racing in the latter stages of its existence and Im sure this is the reason for the poor results.
The final filter we are going to add here is field size, and once again it is a logical inclusion small fields are far more likely to produce a false pace (and therefore a false and unreliable result) than larger ones, and the following fieldsize breakdown supports that theory.
FIELD SIZE
WINS
RUNS
STRIKE%
LSP
LSP%
9 83 181 193
TOTAL
466
1267
36.78
65.04
5.13
At least eight runners in a race is a must have rule for this method, as speed ratings are worth less as the field size diminishes. As Lingfield and Wolverhampton have made the transition to Polytrack, Southwell is the only current allweather course to have a non-Polytrack surface in place, so it is this track which will provide the selections for this method.
So the final rules and results are as follows a) Southwell all-weather races only
b) Horse must be top rated by Topspeed c) Horse must be sent off at shorter odds than 4/1 d) Only consider races 10f or shorter in distance e) Only consider races with at least eight runners
YEAR
WINS
RUNS
STRIKE%
LSP
LSP%
123 90 51 67 43
TOTAL
374
1016
36.81
86.16
8.48