Chapter 5

Introduction
The gambler’s
fallacy and
the ’hot hand’
effect
Conjunction
and
disjunction
fallacies
Base-rate
Judgment under Risk and Uncertainty
neglect
Confirmation
bias
Availability
Manoogian College of Business and Economics, AUA
Overconfidence
Discussion
1 / 30
Introduction
Introduction
The gambler’s
fallacy and • The previous chapter demonstrated how powerful the
the ’hot hand’
effect theory of probabilistic judgment can be.
Conjunction • Few argue that the theory is normatively correct but many
and
disjunction
fallacies
question whether it is descriptively adequate.
Base-rate • This module explores a series of phenomena that suggest
neglect
that the theory of probabilistic judgment does not entirely
Confirmation
bias capture how people actually make probabilistic judgment.
Availability All these phenomena imply that the theory of probabilistic
Overconfidence judgment not descriptively adequate.
Discussion
• In this module, we will also continue our study of the
building blocks of behavioral theory. In particular, we will
continue the discussion of the heuristics-and-biases
program.
2 / 30
Introduction
Introduction
2019-09-13
• The previous chapter demonstrated how powerful the
theory of probabilistic judgment can be.
• Few argue that the theory is normatively correct but many
question whether it is descriptively adequate.
• This module explores a series of phenomena that suggest
Introduction that the theory of probabilistic judgment does not entirely

capture how people actually make probabilistic judgment.
All these phenomena imply that the theory of probabilistic
judgment not descriptively adequate.
• In this module, we will also continue our study of the
building blocks of behavioral theory. In particular, we will
continue the discussion of the heuristics-and-biases
program.
Next time when doing this chapter update the slides on Confirmation bias
and Overconfidence. In David Just’s textbook there is a whole chapter on
these, Chapter 8
Fallacies about Independence
Introduction
The gambler’s • Recall: Outcomes are independent if the occurrence of the one
fallacy and
the ’hot hand’
does not affect the probability that the other one occurs.
effect
• There are in principle two ways in which one might make a
Conjunction
and mistake about independence:
disjunction
fallacies 1. Believing two outcomes are independent when in fact they
Base-rate
are not; and
neglect 2. Believing two outcomes are dependent when in fact they
Confirmation are not.
bias
Availability
• The first fallacy doesn’t have any specific name even though it
Overconfidence
is very common. For example, in financial markets, people often
Discussion
assume that different asset classes behave independently. While
under normal circumstances returns on different asset classes
may be independent, under extreme market movements
(especially market declines) all asset classes may move in the
same direction. Thus, insufficient diversification may be
detrimental.
3 / 30
The gambler’s fallacy
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect • Mistakenly believing two outcomes are dependent when in
Conjunction
and
fact they are not may lead to the gambler’s fallacy.
disjunction
fallacies
• People tend to think that a departures from the average
Base-rate behavior of some system will be corrected in the short
neglect
term.
Confirmation
bias • Example: I flip a fair coin 9 consecutive times and 9
Availability heads come up. Now I am about to flip the coin one more
Overconfidence
time. Which side of the coin would you bet on?
Discussion
• If you picked tails, then you are committing the gambler’s
fallacy.
4 / 30
An example
Introduction
The gambler’s • To see that this is a fallacy, imagine an experiment where

fallacy and
the ’hot hand’ an urn is filled with 5 red and 5 blue balls. Every time a
effect
ball is drawn, it is replaced back, thus this experiment
Conjunction
and involves replacement. As a result, the probability of a
disjunction
fallacies blue or red draw remains the same (1/2). However, people
Base-rate act as if they believe in non-replacement.
neglect
• More specifically, 3 red balls already came out. What is
Confirmation
bias the probability of the next one being blue?
Availability
• A rational person believes that every time a ball is drawn,
Overconfidence
it is replaced back into the urn, thus:
Discussion 4
Pr(B|RRR) = Pr(RRR|B) Pr(B)
Pr(RRR) = (5/10)
(5/10)3
= 12 = 0.5
• An irrational person believes that every time a ball is
drawn it is not replaced back, thus:
Pr(B|RRR) = 5/9×4/8×3/7×5/10
5/10×4/9×3/8 = 60/504
12/72 ≈ 0.71
5 / 30
An example
The gambler’s fallacy and the ’hot hand’ effect

2019-09-13
• To see that this is a fallacy, imagine an experiment where
an urn is filled with 5 red and 5 blue balls. Every time a
involves replacement. As a result, the probability of a
blue or red draw remains the same (1/2). However, people
act as if they believe in non-replacement.
An example • More specifically, 3 red balls already came out. What is

the probability of the next one being blue?
4
Pr(RRR) = (5/10)
(5/10)3
= 12 = 0.5
Pr(B|RRR) = 5/9×4/8×3/7×5/10
5/10×4/9×3/8 = 60/504
12/72 ≈ 0.71
On August 18, 1913, at the casino in Monte Carlo, black came up a

record twenty-six times in succession [in roulette]. . . . [There] was a near-
panicky rush to bet on red, beginning about the time black had come up
a phenomenal fifteen times. In application of the maturity [of chances]
doctrine [the gambler’s fallacy], players doubled and tripled their stakes,
this doctrine leading them to believe after black came up the twentieth
time that there was not a chance in a million of another repeat. In the
end the unusual run enriched the Casino by some millions of francs.
An example

2019-09-13

4
Pr(RRR) = (5/10)
(5/10)3
= 12 = 0.5
Pr(B|RRR) = 5/9×4/8×3/7×5/10
5/10×4/9×3/8 = 60/504
12/72 ≈ 0.71
Carefully note the difference between the following two questions:

(a) You intend to flip a fair coin eight times. What is the probability
that you end up with eight heads?
(b) You have just flipped a fair coin seven times and ended up with
seven heads. What is the probability that when you flip the coin one
last time you will get another heads, meaning that you would have
flipped eight heads in a row?
ANSWER:
(a) P (H&H&H&H&H&H&H&H) = P (H)×P (H)×P (H)×P (H)×
P (H) × P (H) × P (H) × P (H)× = 218 = 256
1
. (b) 12
An example

2019-09-13

4
Pr(RRR) = (5/10)
(5/10)3
= 12 = 0.5
Pr(B|RRR) = 5/9×4/8×3/7×5/10
5/10×4/9×3/8 = 60/504
12/72 ≈ 0.71
• There is a variety of empirical evidence supporting the existence of

the gambler’s fallacy effect. For example, New Jersey’s
pick-three-numbers game is a pari-mutuel betting system; this
means that the fewer people bet on a number, the higher is the
expected payout. It has been found that the amount of money bet
on a particular number falls sharply after the number is drawn, and
only gradually returns to normal after several months (Clotfelter and
Cook, 1993; Terrell, 1994).
Explanation of the Gambler’s Fallacy:
Representativeness and the Law of Small Numbers
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect • The gambler’s fallacy is sometimes explained in terms of
Conjunction
and
the representativeness heuristic.
disjunction
fallacies • You employ the representativeness heuristic when you
Base-rate estimate the probability that some outcome O was the
neglect
result of a given process P by the reference to the degree
Confirmation
bias to which the outcome is representative of that process.
Availability
• If O is highly representative of P , the probability that O
Overconfidence
was a result of P is estimated to be high. Conversely, if O
Discussion
is highly unrepresentative of P , the probability is
estimated to be low.
6 / 30

2019-09-13 • The gambler’s fallacy is sometimes explained in terms of
the representativeness heuristic.
• You employ the representativeness heuristic when you
estimate the probability that some outcome O was the
Explanation of the Gambler’s Fallacy: result of a given process P by the reference to the degree
to which the outcome is representative of that process.
• If O is highly representative of P , the probability that O
Representativeness and the Law of Small was a result of P is estimated to be high. Conversely, if O
is highly unrepresentative of P , the probability is
estimated to be low.
Numbers
• For the first bullet point above: According to the

heuristics-and-biases program, people form judgments by following
heuristics, or rules of thumb, which by and large are functional but
which sometimes lead us percistent biases.
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect • One way to explain the ubiquity of the gambler’s fallacy is
Conjunction
and
to say that people believe in the law of small numbers.
disjunction
fallacies • That is, people exaggerate the degree to which small
Base-rate samples resemble the population from which they are
neglect
drawn. In the case of the coins, the “population” consists
Confirmation
bias of half heads and half tails.
Availability
• A believer in the law of small numbers would exaggerate
Overconfidence
the degree to which a small sample (such as a sequence of
Discussion
nine coin flips) will resemble the population and consist of
half heads and half tails.
7 / 30

2019-09-13 • One way to explain the ubiquity of the gambler’s fallacy is
to say that people believe in the law of small numbers.
• That is, people exaggerate the degree to which small
samples resemble the population from which they are
Explanation of the Gambler’s Fallacy: drawn. In the case of the coins, the “population” consists
of half heads and half tails.
• A believer in the law of small numbers would exaggerate
Representativeness and the Law of Small the degree to which a small sample (such as a sequence of
nine coin flips) will resemble the population and consist of
half heads and half tails.
Numbers
• There is an interesting explanation for the gambler’s fallacy in terms

of evolutionary psychology (Pinker, 1997). It is proposed that in our
past evolutionary environment there was often good reason to
believe that a series of common outcomes would be likely to be
broken at some point. This was particularly true for meteorological
events, like rain or sunshine. Of course, the expected length of the
series would depend on the circumstances, but, just as a cloud
eventually blows past the sun, at some point the probability
becomes higher that in the next time period the sun will come out
again. We shall see that many of the biases that people have are
based on evolutionary adaptations or factors in our past.
The ’Hot Hand’ Effect
Introduction
The gambler’s
fallacy and
the ’hot hand’
• The hot hand effect derives its name from the mistaken
effect
belief among basketball players and fans that a player’s
Conjunction
and chance of hitting a shot is greater following a hit than
disjunction
fallacies following a miss on the previous shot (Gilovich, Vallone
Base-rate and Tversky, 1985).
neglect
Confirmation
• The effect arises when there is uncertainty regarding the
bias distribution of signals, for example, whether a stock price
Availability
will go up or down in any particular time period.
Overconfidence
• Although it appears that this ‘overinference’ is the
Discussion
opposite of the gambler’s fallacy, it is actually a
complementary effect, again involving a misapplication of
the assumption of non-replacement.
8 / 30
Introduction
The gambler’s • To understand what effect a wrong assumption of

fallacy and
the ’hot hand’ non-replacement may have, follow the example of Della
effect
Vigna (2009).
Conjunction
and • There are two types of mutual funds , well-managed and
disjunction
fallacies poorly-managed. An investor is uncertain about the type.
Base-rate 1. The well-managed fund is an urn with 7 UP balls and 3
neglect
DOWN balls, meaning the fund goes up in value 7 times
Confirmation
bias out of 10.
Availability
2. The poorly managed fund is an urn with 3 UP balls and 7
Overconfidence
DOWN balls, meaning it only goes up 3 times out of 10.
Discussion • Before observing any draw of a ball from the urn, the
investor attaches equal probability that the fund is well or
poorly managed. Thus, there is a prior probability of 0.5
that the fund is well managed and a probability of 0.5 that
the fund is poorly managed.
9 / 30
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect
• After observing 3 UP balls, the decision-maker has to
Conjunction determine the probability that the fund is the
and
disjunction well-managed fund.
fallacies
• The rational investor will assume that the balls are
Base-rate
neglect replaced after each draw.
Confirmation
bias • While the irrational investor will assume that the balls are
Availability not replaced.
Overconfidence • Both investors will implement Bayes’ theorem to solve the
Discussion
problem.
• The Bayes formula is again: P (A|B) = P (B|A)P
P (B)
(A)
10 / 30
Introduction
• The rational investor obtains the probability that fund is well
The gambler’s
fallacy and managed as follows:
the ’hot hand’
effect
P (U U U |W ell) × 0.5
Conjunction P (W ell|U U U ) =
and P (U U U |W ell) × 0.5 + P (U U U |P oor) × 0.5
disjunction
fallacies
0.73
= = 0.927
0.73+ 0.33
Base-rate • However, if the irrational agent who behaves according to the law of
neglect
small numbers and assumes that there is no replacement after each
Confirmation
bias draw, the Bayesian expression becomes:
Availability
7/10 × 6/9 × 5/8
Overconfidence
P (W ell|U U U ) = = 0.972
7/10 × 6/9 × 5/8 + 3/10 × 2/9 × 1/8
Discussion • All these mean that after three UP performances
– the rational investor will calculate the probability of an UP
performance as 0.927 × 0.7 + (1 − 0.927) × 0.3 = 0.671
– the irrational (law-of-small-numbers) investor will calculate
the probability of an UP performance as
0.972 × 0.7 + (1 − 0.972) × 0.3 = 0.689
11 / 30
The Gambler’s Fallacy and ’Hot Hand’ Effect: A
Synthesis
Introduction
• Even though, the example above should be sufficient to reconcile the
The gambler’s seemingly contradictory effects of the ‘gambler’s fallacy’ and the ‘hot
fallacy and
the ’hot hand’ hand,’ there have been quite a few approaches to reconcile these two
effect effects.
Conjunction • Barberis, Shleifer and Vishny (1998) demonstrated that
and
disjunction • in the short term, investors may follow the ‘gambler’s
fallacies
fallacy,’ believing that a series of identical signals, like the
Base-rate
neglect
stock price rising, will be followed by a fall (a
Confirmation
‘mean-reverting’ regime). So investors don’t invest in the
bias stock, causing the stock being underpriced. This means
Availability that in the short run, the stock return will demonstrate
Overconfidence positive serial correlation or momentum.
Discussion • in the long run however, investors expect a ‘trending’
regime, whereby the stock is now expected to continue to
rise. This ‘hot hand’ effect causes overreaction, as
investors now overinvest, making the stock overpriced, and
reducing returns, this time demonstrating negative serial
correlation of returns in the long term.
12 / 30
Synthesis
Introduction
The gambler’s • Guryan and Kearney (2008) studied the gambler’s fallacy and the
fallacy and
the ’hot hand’ ’hot hand’ effect in the lottery market.
effect
• They demonstrate that people exhibit both effects by a) not playing
Conjunction
and the winning numbers during subsequent weeks (gambler’s fallacy)
disjunction
fallacies and b) tending to buy more tickets from stores that sold winning
Base-rate tickets in previous weeks (hot hand).
neglect
• Their explanation of this paradoxical combination of the two effects
Confirmation
bias in the lottery market as follows:
Availability – Selection of winning balls involves a random process,
Overconfidence without any intentional human element. Thus, the law of
Discussion small numbers would cause people to to exhibit a
’gambler’s fallacy,’ as they they expect a small sample to
resemble the underlying population.
– However, with stores there could be a human element in
how winning stores are selected, leading to a ‘hot hand.’
13 / 30
Synthesis

2019-09-13
• Guryan and Kearney (2008) studied the gambler’s fallacy and the
’hot hand’ effect in the lottery market.
• They demonstrate that people exhibit both effects by a) not playing
the winning numbers during subsequent weeks (gambler’s fallacy)
and b) tending to buy more tickets from stores that sold winning
tickets in previous weeks (hot hand).
The Gambler’s Fallacy and ’Hot Hand’ Effect: • Their explanation of this paradoxical combination of the two effects
in the lottery market as follows:
– Selection of winning balls involves a random process,
without any intentional human element. Thus, the law of
A Synthesis small numbers would cause people to to exhibit a

’gambler’s fallacy,’ as they they expect a small sample to
resemble the underlying population.
– However, with stores there could be a human element in
how winning stores are selected, leading to a ‘hot hand.’
• For the last sub-bulletpoint: How this human element operates in

this case is open to speculation: it might be that the store is chosen
deliberately by the person buying the winning ticket; or the location
of the winning ticket could be attributable to a corrupt lottery
commissioner, bearing in mind that the winning store owner receives
1% of the prize and thus has an incentive for bribery. Guryan and
Kearney also note that the ‘lucky store’ effect is larger in areas with
more high-school drop-outs, more people living in poverty, and more
elderly. They suggest that this may be caused by cognitive biases.
Conjunction Fallacy
Introduction
The gambler’s • ”A and B” is a conjunction.

fallacy and
the ’hot hand’ • You commit the conjunction fallacy when you
effect
Conjunction
overestimate the probability of a conjunction.
and
disjunction
• Example: The Linda Problem: Linda is 31 years old,
fallacies
single, outspoken and very bright. She majored in
Base-rate
neglect philosophy. As a student, she was deeply concerned with
Confirmation issues of discrimination and social justice and also
bias
participated in anti-nuclear demonstrations.
Availability
Overconfidence (a) What is the probability that Linda is a bank teller?

Discussion
(b) What is the probability that Linda is a bank teller and a
feminist?
• The chance of Linda being a bank teller and a feminist is
sometimes thought to be higher than the chance of her
being a bank teller. This would violate the ”And” Rule.
14 / 30
Conjunction Fallacy
Conjunction and disjunction fallacies

2019-09-13
• ”A and B” is a conjunction.
• You commit the conjunction fallacy when you
philosophy. As a student, she was deeply concerned with
Conjunction Fallacy issues of discrimination and social justice and also

(a) What is the probability that Linda is a bank teller?
feminist?
• Formal solution for the Linda Problem: Let F mean that Linda is
a feminist and B that Linda is a bank teller. Then the probability
that she is both a feminist and a bank teller is
P r(B&F ) = P r(B) × P r(F ) (assuming that the outcomes are
independent). Because P r(F ) ≤ 1 we know that
P r(B) × P r(F ) ≤ P r(B). So whatever the relevant probabilities
involved are, it must be the case that P r(B&F ) ≤ P r(B). Many
people will tell you that Linda is more likely to be a bank teller and
a feminist than she is to be a bank teller.
• By intersecting relatively unrepresentative events (bank teller) with
very representative events (feminist), the conjunction of the two is
considered more probable than the unrepresentative event because
the description is more representative of the combined events. Such
effects may be related to the heuristics and biases that lead to
stereotyping, bigotry, or other potentially undesirable phenomena.
Conjunction Fallacy

2019-09-13

feminist?
Maybe start the section on the Conjunction Fallacy from this example or
from the example on the next note slide.
Example Boeing aircraft A Boeing 747–400 has around 6 million parts.
Suppose that each part is very reliable and only fails with probability
0.000,001. Assuming that failures are independent events, what is the
probability that all parts work?
The probability that any one part works is 0.999,999, so the probability
that all parts work is (0.999, 999)6,000,000 ≈ 0.0025 = 0.25%
A very interesting reference in the book to the ”Jesus nut” that is a nut
on some helicopters that when fail leads to a crush.
Conjunction Fallacy

2019-09-13

feminist?
Lotto 6/49 Many states and countries operate lotteries in which the customer picks
n of m numbers, in any order, where n is considerably smaller than m. In one version
of this lottery, which I will call Lotto 6/49, players circle 6 numbers out of 49 using a
ticket like that in Figure 4.2. The order in which numbers are circled does not matter.
You win the grand prize if all 6 are correct. What is the probability that you win the
Lotto 6/49 any one time you play? Notice that this is similar to picking six consecutive
aces out of a deck with 49 cards, if 6 of those cards are aces.
ANSWER: The answer is: 6/49 × 5/48 × 4/47 × 3/46 × 2/45 × 1/45 = 1/13, 983, 816.
This amounts to about 0.000,000,07. So, if you were to play once a year, on the
average you would win once every 13,983,816 years. If you played once per day, given
that there are 364.25 days in a year, on the average you would win once every 268,920
years. People usually get shocked about how small this number is, in fact they are
overestimating the joint probability that the first number will come out to be right
AND the second number will come out AND...the sixth number will come out right.
Disjunction Fallacy
Introduction
The gambler’s
fallacy and • ”A or B” is a disjunction.
the ’hot hand’
effect • You commit the disjunction fallacy when you
Conjunction
and underestimate the probability of a disjunction.
disjunction
fallacies • Example: A student with a pocket protector might be
Base-rate thought to be more likely to be a physics major than a
neglect
Confirmation
natural sciences major. However, physics is a subset of the
bias natural sciences.
Availability
• Example: Compute the probability of rolling at least one
Overconfidence
six when rolling
Discussion
(a) one die
(b) two dice
(c) three dice
(d) ten dice
15 / 30
Disjunction Fallacy

2019-09-13 • ”A or B” is a disjunction.
• You commit the disjunction fallacy when you
underestimate the probability of a disjunction.
• Example: A student with a pocket protector might be
thought to be more likely to be a physics major than a
Disjunction Fallacy natural sciences major. However, physics is a subset of the

natural sciences.
• Example: Compute the probability of rolling at least one
six when rolling
(a) one die
(b) two dice
(c) three dice
(d) ten dice
1
(a) 6
1 1 1
(b) 6 + 6 − 36 = 11/36 = 0.3056
(c) One (hard) way of answering this question is the following:
P r(A ∨ B ∨ C) = P r(A) + P r(B) + P r(C) − P r(A&B) −
P r(A&C) − P r(B&C) + P r(A&B&C)
= 16 + 16 + 16 − 3 × 36
1
+ 1/216 = 3/6 − 1/12 + 1/216
= 108/216 − 18/216 + 1/396 = 91/216 = 0.421 = 42.1%
But there is an easier way of solving this. Recall that
P r(A ∨ B) = 1 − P r(¬A & ¬B). Thus,
P r(A ∨ B ∨ C) = 1 − P r(¬A & ¬B & ¬C) = 1 − (5/6)3 = 42.1%.
(d) Using the same logic as above: 1 − (5/6)10 = 83.3%
Example
Introduction
The gambler’s
fallacy and
Hiking: You plan to go on a hike in spite of the fact that a
the ’hot hand’
effect
tornado watch is in effect. The national weather service tells
Conjunction
you that for every hour in your area, there is a 30 percent
and
disjunction
chance that a tornado will strike. That is, there is a 30 percent
fallacies chance that a tornado will strike your area between 10 am and
Base-rate
neglect
11 am, a 30 percent chance that a tornado will strike your area
Confirmation between 11 am and noon, and so on.
bias
Availability
(a) What is the probability of a tornado striking your area at
Overconfidence
least once during a two-hour hike?
Discussion (b) What is the probability of a tornado striking your area at
least once during a three-hour hike?
(c) What is the probability of a tornado striking your area at
least once during a ten-hour hike?
16 / 30
Example

2019-09-13
chance that a tornado will strike your area between 10 am and
Example between 11 am and noon, and so on.

(b) What is the probability of a tornado striking your area at
(a) P r(now ∨ in one hour) = 1 − P r(¬now & ¬in one hour) =

1 − (0.7)2 = 0.51 = 51%
(b) P r(now ∨ in one hour ∨ in two hours) =
1 − P r(¬now & ¬in one hour & ¬in two hours) = 1 − (0.7)3 =
0.65.7 = 65.7%
c 1 − (0.7)10 = 97.2%
Example

2019-09-13

The Birthday Problem: Suppose that there are 30 students in your

behavioral economics class.
(a) What is the probability that two students have the same birthday?
(b) What about this class?
Example

2019-09-13

• The birthday problem answer:

(a)Imagine that you line up the 30 students in a row. The first
student can be born on any day of the year, and the probability of
this happening is 365/365; the probability that the second student
does not share a birthday with the first is 364/365; the probability
that the third student does not share a birthday with either of the
first two is 363/365; and so on, until you get to the 30th student:
the probability that this student will not share a birthday with any
of the other 29 students is 336/365. So the probability you are
looking for is 365/365 × 364/365 × ...336/365 ≈ 29.4%. Thus, in a
class this size, the probability that at least two students share a
birthday is quite high: about 70.6 percent. (b) Use the appropriate
number for this part. For example, if there are 9 students in class
then the answer to the problem is:
1 − 364/365 × 363/365 × ... × 357/365 ≈ 1 − 0.91 = 0.9 = 9%
Example

2019-09-13

Example on Flooding Imagine that you live in an area where floods

occur on average every ten years. The probability of a flood in your area
is constant from year to year. You are considering whether to live in your
house for a few more years and save up some money, or whether to move
before you lose everything you own in the next flood.
(a) What is the probability that there will be no floods in your area over
the course of the next two years?
(b) What is the probability that there will be exactly one flood in your
area over the course of the next two years?
(c) What is the probability that there will be at least one flood over the
course of the next two years?
(d) What is the probability that there will be at least one flood over the
course of the next ten years?
Example

2019-09-13

Answers to the flooding example

(a) (9/10)2 = 81.0%
(b) Denote T=flood this year and N=flood next year. Since
P r(T ) = P r(T & N ) + P r(T & ¬N )
⇒ P r(T & ¬N ) = P r(T ) − P r(T & N )
Similarly, P r(N & ¬T ) = P r(N ) − P r(N & T )
Thus,
P r(T & ¬N ∨ N & ¬T ) = P r(T ) + P r(N ) − 2P r(T & N ) =
0.2 − 2 × 0.12 = 0.2 − 0.02 = 0.18 = 18%
(c) 1 − (9/10)2 = 19%
(d) 1 − (9/10)10 = 65.1%
Example

2019-09-13

Example on Terrorism Compute the probability that at least one major

terrorist attack occurs over the course of the next ten years, given that
there are 365.25 days in an average year, if the probability of an attack on
any given day is 0.0001.
Answer 1 − (0.9999)10×365.25 = 30.6%
How Do We Explain the Conjunction and
Disjunction Fallacies?
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect Both the conjunction and disjunction fallacies can be explained
Conjunction in terms of anchoring and adjustment
and
disjunction
fallacies
• People overestimate the probability of conjunctions – and
Base-rate therefore commit the conjunction fallacy – if they use the
neglect
probability of any one conjunct as an anchor and adjust
Confirmation
bias downwards insufficiently.
Availability • They underestimate the probability of disjunctions – and
Overconfidence therefore commit the disjunction fallacy – if they use the
Discussion
probability of any one disjunct as an anchor and adjust
upwards insufficiently.
17 / 30

2019-09-13 Both the conjunction and disjunction fallacies can be explained
in terms of anchoring and adjustment
therefore commit the conjunction fallacy – if they use the
How Do We Explain the Conjunction and probability of any one conjunct as an anchor and adjust
downwards insufficiently.
• They underestimate the probability of disjunctions – and
therefore commit the disjunction fallacy – if they use the
Disjunction Fallacies? probability of any one disjunct as an anchor and adjust

Private jet shopping Suppose you are fortunate (or delusional) enough
to be shopping for a private jet. You have to decide whether to get a jet
with one or two engines. Use p to denote the probability that an engine
fails during any one flight. A “catastrophic engine failure” is an engine
failure that makes the plane unable to fly.
(a) One of the jets you are looking to buy has only one engine. What is
the probability of a catastrophic engine failure during any one flight
with this plane?
(b) Another jet you are looking to buy has two engines, but is unable to
fly with only one functioning engine. Assume that engine failures
are independent events. What is the probability of a catastrophic
engine failure during any one flight with this plane?
(c) Which jet strikes you as safer?
(d) What if the twin-engine jet can fly with only one functioning engine?

2019-09-13 Both the conjunction and disjunction fallacies can be explained
in terms of anchoring and adjustment
therefore commit the conjunction fallacy – if they use the
How Do We Explain the Conjunction and probability of any one conjunct as an anchor and adjust
downwards insufficiently.
• They underestimate the probability of disjunctions – and
therefore commit the disjunction fallacy – if they use the
Disjunction Fallacies? probability of any one disjunct as an anchor and adjust

(a) The probability of a catastrophic engine failure is p.

(b) The probability of a catastrophic engine failure is
1 − (1 − p)2 = 2p − p2 .
(c) The single-engine plane. Notice that when p is small, p2 will be so
small as to be negligible. If so, the twin-engine plan is virtually
twice as likely to experience a catastrophic engine failure as the
single-engine plane!
(d) Now, the probability of a catastrophic engine failure is p2 .
Base-rate neglect
Introduction
The gambler’s
fallacy and
• The base rate is the fraction of some population that has some
the ’hot hand’ characteristic of interest (such as cancer, the common cold, or
effect
the Y chromosome)
Conjunction
and • Base-rate neglect refers to the failure to take the base rate
disjunction
fallacies properly into account in probabilistic judgment.
Base-rate • The Mammogram Example At any one time, 1% of women
neglect
have breast cancer. The test administered is correct in 90% of
Confirmation
bias the cases. That is, if the woman does have cancer, there is a
Availability 90% probability that the test will be positive and a 10%
Overconfidence probability that it will be negative. If the woman does not have
Discussion cancer, there is a 10% probability that the test will be positive
and a 90% probability that it will be negative. Suppose a
woman has a positive test during a routine mammogram
screening. Without knowing any other symptoms, what is the
probability that she has breast cancer?
18 / 30
Base-rate neglect
Base-rate neglect
2019-09-13
characteristic of interest (such as cancer, the common cold, or
the Y chromosome)
• Base-rate neglect refers to the failure to take the base rate
properly into account in probabilistic judgment.
• The Mammogram Example At any one time, 1% of women
Base-rate neglect have breast cancer. The test administered is correct in 90% of
the cases. That is, if the woman does have cancer, there is a
90% probability that the test will be positive and a 10%
probability that it will be negative. If the woman does not have
cancer, there is a 10% probability that the test will be positive
Answer to the mammogram question Denote C = patient has cancer

and P = patient testing positive
From the Baye’s law,
P r(P |C)P r(C)

P r(C|P ) =
P r(P |C)P r(C) + P r(P |¬C)P r(¬C)
0.9 × 0.01
= ≈ 0.08
0.9 × 0.01 + 0.1 × 0.99
Probability of 8% is a bit higher than 1% but much lower than 90% as
many people may think at first. What we are forgetting is that relatively
few people actually have cancer. Out of 1000 people, only about 10 can be
expected to have cancer. Of those, 9 will test positive. Of the 990 women
who do not have cancer, only 10 percent will test positive, but that is still
99 people. So only nine of the 108 people who test positive actually have
cancer, and that is about 8 percent.
Base-rate neglect
Base-rate neglect
2019-09-13
the Y chromosome)
The mamomgram example demonstrates why younger women aren’t rou-

tinely tested fro breast cancer, because of very low base rate, even if they
are tested positive there is a very little chance that it’s a true positive. So
the test is said to be non-diagnostic.
Base-rate neglect
Base-rate neglect
2019-09-13
the Y chromosome)
Down syndrome Example: The probability of having a baby with Down

syndrome increases with the age of the mother. Suppose that the following
is true. For women 34 and younger, about one baby in 1000 is affected
by Down syndrome. For women 35 and older, about one baby in 100
is affected. Women 34 years and younger have about 90 percent of all
babies. What is the probability that a baby with Down syndrome has a
mother who is 34 years or younger?
Answer: Denote Y = mother is yournger than 34 and D=the baby has
down syndrome. We are given P r(D|Y ) = 0.001 and P r(D|¬Y ) =
0.01.P r(Y ) = 0.9 It follows:
P r(D|Y ) × P r(Y )
P r(Y |D) =
P r(D|Y ) × P r(Y ) + P r(D|¬Y ) × P r(¬Y )
0.001 × 0.9 0.0009
= = ≈ 47%
0.001 × 0.9 + 0.01 × 0.1 0.0019
Behavior Detecting Example
Introduction
The gambler’s
fallacy and
the ’hot hand’
• Assuming 60 million people fly every month in the US, and
effect 6 of those are terrorists (0.00001%). Additionally, assume
Conjunction TSA officers are 98% accurate in their assessment of
and
disjunction passengers.
fallacies
(a) What is the probability that a passenger selected at
Base-rate
neglect random is a terrorist and is correctly identified as such by
Confirmation TSA personnel?
bias (b) What is the probability that a passenger selected at
Availability random is not a terrorist but is nevertheless (incorrectly)
Overconfidence identified as a terrorist by TSA personnel?
Discussion (c) What is the probability that a passenger in fact is a
terrorist conditional on having been identified as such by
TSA personnel?
(d) Is this test diagnostic?
19 / 30
Behavior Detecting Example
Base-rate neglect
2019-09-13 • Assuming 60 million people fly every month in the US, and
6 of those are terrorists (0.00001%). Additionally, assume
TSA officers are 98% accurate in their assessment of
passengers.
(a) What is the probability that a passenger selected at
random is a terrorist and is correctly identified as such by
Behavior Detecting Example TSA personnel?

(b) What is the probability that a passenger selected at
random is not a terrorist but is nevertheless (incorrectly)
identified as a terrorist by TSA personnel?
(c) What is the probability that a passenger in fact is a
terrorist conditional on having been identified as such by
TSA personnel?
(d) Is this test diagnostic?
Answer to the behavior detecting example Denote T= the individual

is a terrorist and I = the individual is identified as a terrorist. We are given
that P r(T ) = 0.0000001 and P r(I|T ) = 0.98 and P r(I|¬T ) = 0.02
(a) P r(I & T ) = P r(I|T ) × P r(T ) = 0.98 × 0.0000001 =
0.000000098 = 0.0000098%
(b) P r(I & ¬T ) = P r(I|¬T ) × P r(¬T ) = 0.02 × 0.9999999 =
0.0199999 = 1.9%
(c)
P r(I|T ) × P r(T )
P r(T |I) =
P r(I|T ) × P r(T ) + P r(I|¬T ) × P r(¬T )
0.98 × 0.0000001
= ≈ 0.000005 = 0.0005%
0.98 × 0.0000001 + 0.02 × 0.9999999
(d) No.
Confirmation bias
Introduction
The gambler’s
fallacy and • It is widely believed that that the theory of probability is
the ’hot hand’
effect an adequate normative theory, but its descriptive adequacy
Conjunction is often questioned. The confirmation bias is one of the
and
disjunction
well-known manifestations of this inadequacy.
fallacies • When given new information, rational decision-makers
Base-rate
neglect
must update their believes to objectively reflect the new
Confirmation
information. Bayesian updating is one optimal learning
bias methodology.
Availability • Regardless of the initial believes (priors), when fed with
Overconfidence new information that accurately reflects the true state of
Discussion the world, bayesian learners learn very quickly about the
true state of the world.
• People sometimes refer to this phenomenon as washing
out of the priors.
20 / 30
Confirmation bias
Confirmation bias
2019-09-13 • It is widely believed that that the theory of probability is
an adequate normative theory, but its descriptive adequacy
is often questioned. The confirmation bias is one of the
well-known manifestations of this inadequacy.
• When given new information, rational decision-makers
must update their believes to objectively reflect the new
Confirmation bias information. Bayesian updating is one optimal learning

methodology.
• Regardless of the initial believes (priors), when fed with
new information that accurately reflects the true state of
the world, bayesian learners learn very quickly about the
true state of the world.
• People sometimes refer to this phenomenon as washing
out of the priors.
Start the topic of confirmation bias by asking students how many believe
in global warming?
Confirmation Bias
Introduction
• In real life, unfortunately, people do not in general come to
The gambler’s
fallacy and agree over time.
the ’hot hand’
effect • Sometimes that is because they are exposed to very different
Conjunction evidence: conservatives tend to read conservative newspapers
and
disjunction and blogs while liberals tend to read liberal newspapers and
fallacies blogs. Yet sometimes people have access to the very same
Base-rate evidence presented in the very same way but nevertheless fail to
neglect
agree over time.
Confirmation
bias • Confirmation bias is the tendency of individuals to interpret
Availability evidence as supporting prior beliefs to a greater extent than
Overconfidence warranted.
Discussion • In one classic study, participants who favored or opposed the
death penalty read an article containing ambiguous information
about the advantages and disadvantages of the death penalty.
After reading the article, those who were previously opposed to
the death penalty were even more strongly opposed and those
who favored it were even more in favor.
21 / 30
Confirmation Bias
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect
Conjunction
and
disjunction
fallacies
Base-rate
neglect
Confirmation
bias
Availability
Overconfidence
Discussion
22 / 30
Confirmation Bias Can Explain a Whole Range of
Phenomena
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect
Conjunction
and • Confirmation bias can explain why racist and sexist
disjunction
fallacies stereotypes persist over time.
Base-rate • It can also explain why people gamble.
neglect
Confirmation • Confirmation bias can also explain why so many people

bias
think that they can beat the stock market, in spite of
Availability
evidence that (in the absence of inside information) you
Overconfidence
Discussion
might as well pick stocks randomly.
23 / 30
What Causes Confirmation Bias?
Introduction
The gambler’s
fallacy and
the ’hot hand’ Psychological research suggests that confirmation bias is due to
effect
a number of different factors.
Conjunction
and
disjunction
1. People sometimes fail to notice evidence that goes against
fallacies
their beliefs, whereas they quickly pick up on evidence
Base-rate
neglect
that supports them.
Confirmation 2. When the evidence is vague or ambiguous, and therefore
bias
Availability
admits of multiple interpretations, people tend to interpret
Overconfidence
it in such a way that it supports their beliefs.
Discussion 3. People tend to apply a much higher standard of proof to
evidence contradicting their beliefs than to evidence
supporting them.
24 / 30
Availability
Introduction
The gambler’s
fallacy and
• People are often lousy at estimating probabilities of rare
the ’hot hand’
effect
events. They overestimate the probability of dying in a
Conjunction plane crash, or in pregnancy, or suffering from violent
and
disjunction crime. They also tend to overestimate the probability of
fallacies
winning a lottery.
Base-rate
neglect • People tend to believe that events are more frequent or
Confirmation
bias
more probable if examples of it are easier to remember.
Availability • Availability is the ease with which information can be
Overconfidence brought to mind.
Discussion • The availability heuristic says that we can treat X as
more likely than Y if X comes to mind more easily than Y .
• Like other heuristics, the availability heuristic may be
functional but occasionally leads us astray.
25 / 30
Availability
Availability
2019-09-13
plane crash, or in pregnancy, or suffering from violent
crime. They also tend to overestimate the probability of
winning a lottery.
• People tend to believe that events are more frequent or
Availability more probable if examples of it are easier to remember.

• Availability is the ease with which information can be
brought to mind.
• The availability heuristic says that we can treat X as
For the last bullet point on the previous slide give this example: Suppose,
for instance, that you happen to come across an alligator while walking to
work. The chances are that images of alligators attacking other animals
(including humans) will come to mind more easily than images of alligators
acting cute and cuddly. If so, the availability heuristic tells you to assume
the alligator is likely to be dangerous, which is obviously a helpful assump-
tion under the circumstances. However, the availability heuristic (being a
simple rule of thumb) can sometimes lead you astray, as in the case of the
children with headaches. Thus, like anchoring and adjustment, availability
can lead to bias.
Availability
Availability
2019-09-13
winning a lottery.

brought to mind.
CT scans In some populations, brain tumors in children are rare: the base
rate is only about 1/10,000. A child with a tumor is very likely to have
occasional headaches: 99 out of 100 do. But there are many other reasons
a child can have a headache: of those who do not have a tumor 1 in 10
have occasional headaches.
(a) Given that a child has occasional headaches (H), what is the
probability that he or she has a brain tumor (T)?
(b) Let us say that among children with headaches, 999/1000 will
ultimately be fine (F). Suppose that a physician using a simple test
can correctly determine whether the child will be fine or not in
95/100 of children with headaches. Given that the doctor after
performing the test gives the patient a green light (G), what is the
probability that the child really will be fine?
Availability
Availability
2019-09-13
winning a lottery.

brought to mind.
Solution:
P r(H|T )×P r(T )
(a) P r(T |H) = P r(H|T )×P r(T )+P r(H|¬T )×P r( negT )
99/100×1/10,000
99/100×1/10,000+1/10×9999/10,0000 ≈ 0.001
P r(G|F )×P r(F )

(b) P r(F |G) = P r(G|F )×P r(F )+P r(G|¬F )×P r(¬F )
95/100×999/1,000
95/100×999/1,000+5/100×1/1,0000 ≈ 0.99996
Availability
Introduction
Availability heuristic can explain a variety of phenomena, including
The gambler’s
fallacy and • Why people think violent crime is more common than other
the ’hot hand’
effect
kinds of crime.
Conjunction
• Why fears of airplane crashes, nuclear meltdowns, and terrorist
and
disjunction
attacks tend to increase dramatically shortly after such events
fallacies occur, for the obvious reason that they come to mind
Base-rate particularly easily then.
neglect
• Why people tend to oppose free international trade.
Confirmation
bias • Why anti-vaccination sentiments are so strong in spite of
Availability overwhelming evidence that vaccines are safe and effective.
Overconfidence • Why people repeat dangerous behaviors.
Discussion Availability bias also helps explain the base-rate fallacy. In the case of
cancer, the cases of true positives (people who were correctly
diagnosed with cancer) are more likely to come to mind than cases of
false positives (people who were incorrectly diagnosed with cancer).
Thus, we are likely to overestimate the probability of cancer when
tested positive.
26 / 30
Availability
Availability Availability heuristic can explain a variety of phenomena, including
2019-09-13
• Why people think violent crime is more common than other
kinds of crime.
• Why fears of airplane crashes, nuclear meltdowns, and terrorist
attacks tend to increase dramatically shortly after such events
occur, for the obvious reason that they come to mind
particularly easily then.
Availability • Why people tend to oppose free international trade.

• Why anti-vaccination sentiments are so strong in spite of
overwhelming evidence that vaccines are safe and effective.
• Why people repeat dangerous behaviors.
Availability bias also helps explain the base-rate fallacy. In the case of
cancer, the cases of true positives (people who were correctly
diagnosed with cancer) are more likely to come to mind than cases of
false positives (people who were incorrectly diagnosed with cancer).
Thus, we are likely to overestimate the probability of cancer when
tested positive.
Causes of death According to the World Health Organization, the leading

causes of death in the world are ischaemic heart disease, stroke, lower
respiratory infections, and chronic obstructive lung disease. This makes
the leading causes of death quite different from the leading sources of
fear. The effect is that people spend relatively much time thinking about
threats they can do little about (terrorist groups and epidemic diseases
across the world, for example) and relatively little time thinking about
things they can (such as cardiovascular health, which can be improved by
exercise and diet). Availability can explain why people overestimate the
danger posed to them by the former and underestimate that by the latter.
But a no less interesting question is this: Can the power of availability and
other heuristics be harnessed to encourage people to think more about
things such as cardiovascular health?
Overconfidence
Introduction
The gambler’s • The probabilities that we assign to uncertain outcomes

fallacy and
the ’hot hand’ manifest our confidence: statements concerning how
effect
certain you are that various things will happen.
Conjunction
and
disjunction
• When assessing statements of confidence, behavioral
fallacies
economists talk about calibration.
Base-rate
neglect • Formally speaking, you are perfectly calibrated if, over the
Confirmation long run, for all propositions assigned the same probability,
bias
Availability
the proportion of the propositions that are true equals the
Overconfidence
probability you assigned.
Discussion • One of the most persistent findings in the literature is that
people tend to exhibit overconfidence. The certainty that
people (including experts) express in their judgments tends
to exceed the frequency with which those judgments are
correct.
27 / 30
Overconfidence
Overconfidence
2019-09-13
• The probabilities that we assign to uncertain outcomes
manifest our confidence: statements concerning how
certain you are that various things will happen.
• When assessing statements of confidence, behavioral
economists talk about calibration.
• Formally speaking, you are perfectly calibrated if, over the
Overconfidence long run, for all propositions assigned the same probability,
the proportion of the propositions that are true equals the
probability you assigned.
• One of the most persistent findings in the literature is that
people tend to exhibit overconfidence. The certainty that
people (including experts) express in their judgments tends
to exceed the frequency with which those judgments are
correct.
In early studies, researchers asked undergraduates questions of the type

“Absinthe is (a) a liqueur or (b) a precious stone?” and invited them to
judge how confident they were that their answer was right. Participants
who indicated that they were 100 percent certain that their answers were
right were on the average correct 70–80 percent of the time.
Meteorology Evidence suggests meteorologists are well calibrated and

therefore an exception to the rule. This will strike many people as lit-
erally unbelievable. What heuristic might cause them to underestimate
meteorologists’ ability to offer calibrated predictions?
Answer If people are more likely to remember cases when meteorologists’
predictions were off, which is likely, the availability heuristic will cause
people to think meteorologists are more poorly calibrated than they really
are.
Overconfidence
Introduction
The gambler’s
fallacy and
the ’hot hand’
effect
Conjunction • Studies indicate that overconfidence increases with

and
disjunction confidence, and therefore is most extreme when confidence
fallacies
is high.
Base-rate
neglect • Overconfidence also increases with the difficulty of the
Confirmation
bias judgment task. The more challenging the task is, the more
Availability likely a judge is to be overconfident.
Overconfidence • Interestingly, overconfidence does not in general seem to
Discussion
decrease when people become more knowledgeable.
28 / 30
Overconfidence
Overconfidence
2019-09-13 • Studies indicate that overconfidence increases with
confidence, and therefore is most extreme when confidence
is high.
• Overconfidence also increases with the difficulty of the
Overconfidence judgment task. The more challenging the task is, the more
likely a judge is to be overconfident.
• Interestingly, overconfidence does not in general seem to
decrease when people become more knowledgeable.
In one famous study, the researcher asked participants in his study ques-
tions about the behaviors, attitudes, and interests of a real patient. As
participants received more and more information about the patient’s life,
they assigned more and more confidence to their answers. Yet, their ac-
curacy barely increased at all. Incidentally, the clinical psychologists who
participated in the study – a majority of whom had PhDs – were no more
accurate and no less confident than psychology graduate students and
advanced undergraduates. The more educated people were actually more
overconfident than the less educated people. It may well be true, as Alexan-
der Pope said, that “[a] little learning is a dangerous thing” – and a lot of
learning too.
What Causes Overconfidence?
Introduction
The gambler’s
• Many of our judgments are not repetitive, and we do not receive
fallacy and
the ’hot hand’
feedback that is regular, prompt, and unambiguous.
effect • Furthermore, even in the presence of outcome feedback,
Conjunction
and
learning from experience is more difficult than one might think.
disjunction Confirmation bias makes us overweight evidence that confirms
fallacies
our predictions and underweight evidence that disconfirms them.
Base-rate
neglect • Availability bias does not help either. If the image of a situation
Confirmation where you were right when others were spectacularly wrong
bias
comes to mind easily and often, you might end up
Availability
overestimating the probability that that sort of thing will
Overconfidence
happen again.
Discussion
• The overconfidence phenomenon receives indirect support from
research on competence; many studies suggest that people
overestimate their competence in various practical tasks. The
vast majority of drivers – in some studies, more than 90 percent
– say that they are better than the median driver.
29 / 30
What Causes Overconfidence?
Overconfidence
2019-09-13
• Many of our judgments are not repetitive, and we do not receive
feedback that is regular, prompt, and unambiguous.
• Furthermore, even in the presence of outcome feedback,
learning from experience is more difficult than one might think.
Confirmation bias makes us overweight evidence that confirms
our predictions and underweight evidence that disconfirms them.
• Availability bias does not help either. If the image of a situation
What Causes Overconfidence? where you were right when others were spectacularly wrong
comes to mind easily and often, you might end up
overestimating the probability that that sort of thing will
happen again.
• The overconfidence phenomenon receives indirect support from
research on competence; many studies suggest that people
overestimate their competence in various practical tasks. The
vast majority of drivers – in some studies, more than 90 percent
– say that they are better than the median driver.
In one fascinating study, undergraduates whose test scores in grammar and logic put
them in the bottom 25 percent of a group of peers, on the mean estimated that they
were well above average. Even more surprising, perhaps, when participants received
more information about their relative performance in the tests (by being asked to grade
those of other participants), the strongest students became more calibrated, while the
weakest students, if anything, became less calibrated. The results suggest the least
competent are at a double disadvantage, in that their incompetence causes both poor
performance and an inability to recognize their performance as poor: the cognitive
skill required to perform a difficult mental task may well be tightly tied up with the
metacognitive skill to assess the quality of our performance. This Dunning–Kruger
effect may or may have been what comedian Ricky Gervais was getting at when he said:
“When you are dead, you do not know you are dead. It is only painful for others. The
same applies when you are stupid.”
Discussion
Introduction
The gambler’s
• This chapter has explored phenomena that appear
fallacy and
the ’hot hand’
inconsistent with the theory of probabilistic judgment that
effect
we learned about in Chapter 4.
Conjunction
and
• While the theory of probability was never designed to
disjunction
fallacies capture the precise cognitive processes people use when
Base-rate forming judgments, there appears to be a wide range of
neglect
circumstances under which people’s intuitive probability
Confirmation
bias judgments differ substantially, systematically, and
Availability predictably from the demands of the theory.
Overconfidence • We have also discussed some of the theoretical tools used
Discussion by behavioral economists to capture the manner in which
people actually make judgments. Thus, we explored
further aspects of the heuristics-and-biases program, which
is one prominent effort to develop a descriptively adequate
theory of probabilistic judgment.
30 / 30

Chapter 5

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 5

Uploaded by

Copyright:

Available Formats

Introduction

Introduction that the theory of probabilistic judgment does not entirely

The gambler’s • To see that this is a fallacy, imagine an experiment where

The gambler’s fallacy and the ’hot hand’ effect

An example • More specifically, 3 red balls already came out. What is

On August 18, 1913, at the casino in Monte Carlo, black came up a

The gambler’s fallacy and the ’hot hand’ effect

An example • More specifically, 3 red balls already came out. What is

Carefully note the difference between the following two questions:

The gambler’s fallacy and the ’hot hand’ effect

An example • More specifically, 3 red balls already came out. What is

• There is a variety of empirical evidence supporting the existence of

The gambler’s fallacy and the ’hot hand’ effect

• For the first bullet point above: According to the

The gambler’s fallacy and the ’hot hand’ effect

• There is an interesting explanation for the gambler’s fallacy in terms

The gambler’s • To understand what effect a wrong assumption of

The gambler’s fallacy and the ’hot hand’ effect

A Synthesis small numbers would cause people to to exhibit a

• For the last sub-bulletpoint: How this human element operates in

The gambler’s • ”A and B” is a conjunction.

Overconfidence (a) What is the probability that Linda is a bank teller?

Conjunction and disjunction fallacies

Conjunction Fallacy issues of discrimination and social justice and also

Conjunction and disjunction fallacies

Conjunction Fallacy issues of discrimination and social justice and also

Conjunction and disjunction fallacies

Conjunction Fallacy issues of discrimination and social justice and also

Conjunction and disjunction fallacies

Disjunction Fallacy natural sciences major. However, physics is a subset of the

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

(a) P r(now ∨ in one hour) = 1 − P r(¬now & ¬in one hour) =

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

The Birthday Problem: Suppose that there are 30 students in your

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

• The birthday problem answer:

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

Example on Flooding Imagine that you live in an area where floods

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

Answers to the flooding example

Conjunction and disjunction fallacies

Example between 11 am and noon, and so on.

Example on Terrorism Compute the probability that at least one major

Conjunction and disjunction fallacies

Disjunction Fallacies? probability of any one disjunct as an anchor and adjust

Conjunction and disjunction fallacies

Disjunction Fallacies? probability of any one disjunct as an anchor and adjust

(a) The probability of a catastrophic engine failure is p.

Answer to the mammogram question Denote C = patient has cancer

P r(P |C)P r(C)

The mamomgram example demonstrates why younger women aren’t rou-

Down syndrome Example: The probability of having a baby with Down

Behavior Detecting Example TSA personnel?

Answer to the behavior detecting example Denote T= the individual

Confirmation bias information. Bayesian updating is one optimal learning

Confirmation • Confirmation bias can also explain why so many people

Availability more probable if examples of it are easier to remember.

Availability more probable if examples of it are easier to remember.