Chan, Ho Fai; Savage, David; Torgler, Benno

Working Paper
Sport as a behavioral economics lab

CREMA Working Paper, No. 2021-20

Provided in Cooperation with:

CREMA - Center for Research in Economics, Management and the Arts, Zürich

Suggested Citation: Chan, Ho Fai; Savage, David; Torgler, Benno (2021) : Sport as a behavioral
economics lab, CREMA Working Paper, No. 2021-20, Center for Research in Economics,
Management and the Arts (CREMA), Zürich

This Version is available at:

https://hdl.handle.net/10419/234635

Standard-Nutzungsbedingungen: Terms of use:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your
Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial
Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them
machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise
use the documents in public.
Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen
(insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open
gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you
genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated
licence.
 Center for Research in Economics, Management and the Arts

Raumplanung:
Rückzonungen sollen
Einzonungen ermöglichen
Sport as a Behavioral Economics Lab

René L. Frey

Artikel erschienen in Basellandschaftliche Zeitung, 28. November 2012, S. 30,

aufgrund des Referats «Mehrwertabschöpfung: Eine politisch-ökonomische Analyse»,
gehalten am 1. November 2012 in Zürich im Rahmen des «Forums Raumwissenschaften»,
Universität Zürich und CUREM

Beiträge zur aktuellen Wirtschaftspolitik No. 2012-04

Working Paper No. 2021-20

CREMA Gellertstrasse 18 CH-4052 Basel www.crema-research.ch

CREMA Südstrasse 11 CH - 8008 Zürich www.crema-research.ch
 Sport as a Behavioral Economics Lab

Ho Fai Chan1,2, David A. Savage2,3, and Benno Torgler1,2,4

School of Economics and Finance, Queensland University of Technology, Gardens Point, 2 George St,
Brisbane, QLD 4001, Australia.
2
Centre for Behavioural Economics, Society and Technology, Queensland University of Technology,
Brisbane, Australia.
3
Newcastle Business School, University of Newcastle, 409 Hunter Street, Newcastle, NSW, 2200,
Australia.
4
CREMA—Center for Research in Economics, Management and the Arts, Switzerland

Contribution prepared for the forthcoming book: Hannah J. R. Altman, Morris Altman, and
Benno Torgler (Eds.) (2021). Behavioural Sports Economics. New York: Routledge

ABSTRACT
Sporting events can be seen as controlled, real-world, miniature laboratory environments, approaching
the idea of “holding other things equal” when exploring the implications of decisions, incentives, and
constraints in a competitive setting (Goff and Tollison 1990, Torgler 2009). Thus, a growing number
of studies have used sports data to study decision making questions that have guided behavioural
economics literature. Creative application of sports data can offer insights into behavioural aspects
with implications beyond just sports. In this chapter, we will discuss the methodological advantages of
seeing sport as a behavioural economics lab, concentrating on the settings, concepts, biases, and
challenging areas. Beyond that, we will discuss questions that have not yet been analysed, offering
ideas for future studies using sports data. We will further reflect on how AI has evolved; focusing, for
example, on chess, which provides insights into the mechanism and machinery of decision-making.


For helpful comments and suggestions thanks are due to the Jillian Cortese and Alison Macintyre.

1
Introduction
Sports economics, or the study of sport using economic theory, has been with us for some time

and has generated interesting insights into several different topics. Of course, one principle that

sits behind all this analysis is the methodology or approach through which we try and apply

our thinking to sports data. The behavioural economics revolution of recent years has given

economists an additional lens through which they can examine the world around them – this is

no different for the analysis of sports data. We have seen an increase in the number of papers

that have utilised sports data to gain insight or understanding into individual, group, or other

types of behaviour. We have used economics to explore sports data and aspects of the real-

world that are often very difficult to access or observe, such as performance and incentives.

These aspects are much easier to study in the world of sports, as sport data is all-encompassing

and available in almost every type of sport. To discuss how behavioural economics influenced

research with sports data, we will first discuss setting conditions that provide insights to the

advantage of sports data. Next, we will discuss a set of concepts that have been explored or are

worth exploring in more detail. These concepts were chosen ad hoc and therefore are not a

complete list of valuable concepts. Moreover, as the list is relatively substantial, we do not

provide a full list of available studies, but rather a set of studies that clarify the area of

exploration and those where we see further possibilities and avenues. We therefore apologize

in advance if some scholars feel that their work should have been mentioned under these

concepts. This process was like wandering through a field of gems with only a small bag in

which to collect them – you obviously cannot collect them all, nor could you even be sure of

total gems available in the field. We cherrypicked the gems that we felt best represented the

discussion we were trying to build. We finish the article by discussing some areas that have

influenced or were influenced by behavioural economics, focusing specifically on biases and

linking behavioural economics with AI, which will become a more dominant area in the future

2
(Schmidt 2020, Torgler 2020). Looking at chess, for example, provides insights into the

mechanism and machinery of decision-making and how such mechanisms can be programmed

to derive more realistic cognitive architectures (Torgler 2021a). We will clarify why past

insights are valuable from a behavioural economics perspective, and what we can expect in

terms of future perspectives.

Settings

In this section we will identify interesting settings that allow us to understand how sports can

be seen as a “real-world laboratory” (Goff and Tollison 1990, Torgler 2009, Kahn 2000) where

we can test behavioural questions in a relatively controlled high-stakes environment, and where

information is transparently available for outsiders to explore (rather than hidden as is often

the case in the labour market). However, as controllability is influenced by and subject to

various elements and aspects, we try to provide a set of examples.

Rule Changes

With game rule changes, we are approaching conditions of a natural experiment (Chan, Savage,

and Torgler 2019); accordingly, scholars have focused on those rule changes since almost the

beginning of sports economics. A frequent example that appears in the literature examines how

changes in the number of referees (change in enforcement ability) affect players’ behaviour in

environments such as basketball (McCormick and Tollison 1984), ice hockey (Levitt 2002), or

soccer (Witt 2005). There are historically fascinating examples in which randomization of

referees was introduced, such as in the 1998-99 National Hockey League (NHL) season when

the assignment of one or two referees to a match was done randomly (Levitt 2002). Chan et al.

(2019) also explored not just the one rule change, but repeated changes that led to a return to

the original condition. This allowed readers to see not just a first adaption process, but also re-

3
adaption. Rule changes are particularly interesting from a behavioural economics perspective

because a key intention of sports rule change is to change the behaviour within the game (Elias

and Dunning 1966, Chan et al. 2019). Such changes can affect aspects such as emotionality,

predictability, uncertainty, cooperation, or the psychology of an athlete in general. The beauty

is not only in the uncertainty created, but also in understanding the implications of observing

behaviour in a relatively close micro-environment that has only limited spill-over effects from

other areas. Uncertainty in other natural environments beyond sports is usually more complex

and less controlled.

Those empirical explorations are important as behavioural economics needs to get the

psychology as well as the economics right, and therefore how humans respond to incentive

changes and changes in their constraints and opportunities. Thus, linking economics and

psychology together in the way behavioural economics can help to gain insights into the

mechanisms of behavioural change using data beyond experimentation. If we are keen to

understand how a billiard player adapts to changes it is not enough to just assume she/he is

familiar with the physics and mathematics of producing perfect shots. We need to understand

how those players act, learn, and adapt. Herbert Simon has clearly won that debate against

Milton Friedman: training them requires answering “why” questions. Think about heuristics

(and potential biases) and decision-making with a mind-set that tries to understand how the

actual world of sports works and evolves with all its underlying mechanisms. This means that

rule changes can be seen as some sort of a resilience test that helps us to understand those

mechanisms better.

Institutional Factors

There are interesting links between the way that institutions are set up and how they influence

human behaviour that is often overlooked, which borders on ironic given that most are little

4
more than social norms that are sufficiently long-lived, becoming an institution by formalising

a set of rules to govern how people should act and behave. The process of becoming an

institution describes a feedback mechanism between individuals and groups on shared social

norms and the mores of acceptable behaviour (Savage 2019). However, it also signifies that it

takes a significant length of time and stability for a norm to become institutionalised, implying

that they do not change quickly and are likely well out of step with the norms of a quickly

evolving society. This is no different for sporting institutions, especially those that have been

functional for centuries, as the values and norms of the society that built the institution may

have moved on, leaving the institution struggling to remain relevant or effective.

Institutions define incentives and, therefore, behaviours. A good example is a study

provided by Duggan and Levitt (2002). The ancient and noble sport of Sumo Wrestling was

considered to be above the cheating and corruption that plagued nearly all others. However,

the authors revealed that is not the case when looking at how the structure of the competition

was examined through a more behavioural lens. The analysis focused on the how the sport’s

unique competition structure incentivised collusive behaviours between opponents at a kink

point in the tournament structure (when moving from seven to eight wins) where the payoff –

a promotion of rank – significantly increased for one competitor if they were successful (eight

wins). The eighth win has four times the value of a typical win. The analysis showed that on

any other bout when there was no such increased incentive to win, the outcomes were within

the expected win/loss ratios, but on the rounds where one player would receive the promotion,

the probability of their winning escalated dramatically. Effectively they demonstrated that the

opposing player was willing to collude by throwing the bout in order to ensure the promotion

(moving up a single spot would be worth $3000), but they also found that when these

competitors next faced each other the ‘favour’ was returned as the promoted player would most

likely lose the match. Such collusion shows that this cannot be explained purely by an athlete

5
effort story. Match rigging also increased as the tournament ended. The results show that

wrestlers on the bubble on day 15 are victorious 25 percent more often than would be expected.

On the other hand, excess winning likelihoods disappeared in tournaments with a high level of

media scrutiny. Their findings are interesting for several reasons. It shows that corruption was

present in even a sport thought to be above such activities, but more importantly it

demonstrated that the incentive structure put in place by the sporting institution had a

significant impact on the attitudes and behaviours of players i.e., it inadvertently created a focal

point for player collusion at a specific point in every tournament. In addition, their results show

that the cost of detection also matters. Increasing media scrutiny can help to reduce collusive

behaviour, while an interesting avenue for further investigation would be whether, and to what

extent, sports disciplines have barriers to entry in the market that affect corruption or

innovation by lacking competitive pressure.

Institutions have been developed to help codify and regulate international sport with

the creation of professional associations and presence of referees as the on-field arbiter of the

rules. These institutions ensure that the negative aspects of competitive sports, i.e., aggression,

conflict, and violence do not degenerate into a free-for-all, and that they mirror the acceptable

behaviours we observe within our modern societies (Howell 1975, Cooper 1989, Riordan

1993). In line with this, Caruso et al. (2017) explored the relationship between international

football competitions and conflict across several international football tournaments including

the FIFA World Cups, Olympic Games, Champions Cups, and Under 20’s World Cups

between 1994 and 2014 to analyse the impact that national differences (identity) had on match

aggression and conflict. They conclude that while the impact of any one referee on enforcing

the institution’s position is not clear, the analysis of several decades of competitions across

numerous international tournaments enabled them to ascertain the true impact of the referees

controlling conflict during those matches. Furthermore, they demonstrate that FIFA, the

6
institution governing football, was able to have a positive impact on behaviour by limiting or

directing social violence through the control of referees. By examining aggression and violence

through the awarding of red or yellow cards and in-game fouls for minor indiscretions, they

demonstrated that match referees had a significant institutional impact on the on-field

transgressions that negated the significance of virtually all the national identity variables. This

result demonstrates that when an institution’s rules and expectations are well aligned and are

consistently applied and updated, they can have a significant positive impact on behavioural

outcomes. The authors also warn “that if an institutional approach is adopted the evidence

suggests that it needs a certain degree of flexibility and prescience, rather than a rule based

reactive rigidity” (Caruso et al. 2017, p. 538).

Disruptions

The sports environment, team sports in particular, provides an interesting environment to

explore organizational disruptions. In-season changes, such as dismissing head-coaches, offers

a natural avenue of investigation, allowing us to understand the implications of major

management changes. Such head coach changes are frequent enough to allow detailed

exploration. Van Ours and van Tuijl (2016) report mean in-season head-coach changes

between 4.2 (Netherlands Eredivisie) and 8.4% (Italy Serie A) for seven European football

leagues, covering 14 seasons, starting with the 2000/2001 season. Their results indicate that a

replacement improves performance.

Another interesting aspect is understanding what happens if a team recruits a highly

talented athlete; a star. Sports allows us to explore and track what happens individually and

teamwise over time. It can therefore provide better insights into dynamics that are difficult to

observe in the normal labour force. Evidence regarding CEOs indicates that top performers

achieve success only for a while, fading out quickly and leading to a sharp decline in teams’

7
functioning (Groysberg, Nanda and Nohria 2004). Those insights indicate that companies

cannot gain from a competitive advantage by hiring stars, considering that they do not stay

with organizations for a long time (Groysberg, Nanda and Nohria 2004). Having detailed data

on how players interact with each other on the field (e.g., who interacts with whom throughout

the game or even during training sessions) allows us to observe and understand better how stars

are integrated into the team, how that integration evolves over time, and whether and to what

extent an integration is affected by the actual structure of the team itself (e.g., level of

heterogeneity across a number of factors such as age, experience, loyalty to the team, salary,

etc.). Future studies can also explore in more detail how technological disruptions such as Big

Data, AI, or quantum computing affect the game itself (Torgler 2020).

Pressure

Experiments involving pressure and stress have been directly linked to the breakdown in

judgment, rational decision making, and the generation of mistakes via an individual’s inability

to correctly weight options (Wright 1974). This can result in inefficient or poor outcomes as

individuals fail to correctly scan alternative options (Keinan 1987). As pressure increases,

individuals are less able to make rational choices, resulting in a greater number of irrational

choices (Meichenbaum 2007). However, it remains unclear exactly how much additional

pressure is required to detrimentally affect decisions (Jamal 1984), and part of this problem

derives from differences in the underlying model being used in the stress function. For

example, Sullivan and Bhagat (1992) outline four of the most common models, which

included: 1) higher levels of performance require at least some moderate level of stress; 2) a

positively correlated relationship such that only through high stress could high performance be

achieved; 3) a negatively correlated relationship where high stress results in low performance

levels; 4) and finally that stress and performance are totally unrelated. Additional issues arise

8
when non-linear models like an inverted ‘U’ shape function (Allen et al. 1982, Meglino 1977,

Yerkes and Dodson 1908) are adopted over linear ones – where lower levels of stress may

actually aid in performance, but once the turning point (threshold) has been reached any

additional stress is detrimental to performance (see e.g., Baumeister 1984, Baumeister and

Showers 1986).

Sports allow us to pin-point the actual pressure experienced (magnitude and direction

of stress) relatively precisely. Savage and Torgler (2012) explored the impact of different stress

factors on elite athletes during penalty shoot-outs at the FIFA World Cup and the UEFA Euro

Cup competitions between 1978 and 2008. They found that predicable, anticipated, and

experienced stress factors (routinely experienced stress determinants), such as crowd size

(noise) or game level, have no impact on performance. However, the less anticipated stressors

such as final shots to win/lose appeared to have significant impacts on likelihood of success.

A large positive difference promotes performance (positive stress) and improves the

probability of a successful shot by about 17%, while a negative difference reduces performance

(negative stress) and decreases the probability of a successful shot by about 45%. This indicates

a substantial and asymmetric effect for top athletes, which means that they also respond

differently to detrimental incentive effects (high rewards or the threat of severe failure).

Krumer (2020) shows that penalty kicks are not just a “lottery”. Exploring the probability of

winning a shoot-out by looking at teams from different divisions indicates that higher-ranked

soccer teams perform better in penalty kicks.

The penalty kick environment has also been used to understand the effect of a

supportive audience on performance. Dohmen (2008) found that players in the German

Bundesliga (seasons between 1963 and 2004) were more likely to “choke” when playing in

front of their home audience. Thus, empirical evidence using sports data are useful as such data

can help to discriminate between theories. For example, in this case one can discriminate

9
between a social facilitation or social support hypothesis that suggests that performance is

boosted by a friendly environment, and a social pressure hypothesis which argues that it can

impair performance (Dohmen 2008) via a higher psychological pressure due to higher

expectations. As Butler and Baumeister (1998) stress “it may be more painful to have friends

and family see one fall flat on one’s face. A supportive audience could conceivably increase

pressure, concern, and self-consciousness, which in principle could have a detrimental effect

on performance” (p. 1213). Applying an experimental approach, they observe that choking

under pressure around a supportive audience is found in skill-based tasks but not in easy tasks,

despite finding supportive audiences to be more helpful and less stressful. This led the authors

to conclude that people were not aware of the debilitating effect of supportive audiences.

Beyond such experiments, sports data provides a controlled setting in the real world to

explore whether or not – and in what conditions – choking under pressure matters. Dohmen

(2008, p. 652) stresses that penalty kicks are not free of problems. Only a selected group of

players is explored; therefore, a selection effect may lead to a lower bound estimate, as those

who are better able to cope with stress are more likely to be selected for such penalty kicks.

Using penalty kicks in World Cups after a draw would help to reduce such a problem as it may

increase the distributions of ability to choke (or not) under pressure (despite some ex-ante

training). Dohmen also stresses that for Bundesliga data, the stakes might not be high enough

to observe significant choking, which also means that such results would be lower bound

estimates (compared to other settings). Using World Cup or UEFA European Cup data would

mean exploring players in a higher stress environment:

On 17 July 1994, at the Los Angeles Rose Bowl, Brazil attempted to secure its fourth Federation

Internationale de Football Association (FIFA) World Cup trophy, in probably one of the most

memorable shootouts in World Cup history. One of Italy’s greatest ever players and a shining

light of the tournament, Robert Baggio, took what was to be the final shot of the US World

Cup. Baggio placed the ball on the spot, while Taffarel, the Brazilian goal keeper, took his

10
 position on the line in front of 94 000 spectators. The fascinating aspect of such a ‘high

pressure’ situation is the fact that after 4 years of preparation, several matches before this final,

120 min of game time, and eight prior penalty attempts, one single shot held the match outcome

in the balance. If Baggio misses then Italy loses the greatest prize of all in football, namely the

World Cup; if he is successful Italy still can retain a glimmer of hope of being world champions.

As many readers may know, Baggio’s shot not only missed but it soared metres over the

crossbar, which meant that Italy lost the tournament and Brazil became the 1994 World Cup

champions (Savage and Torgler 2012, pp. 2423-2424).

Beyond soccer, one can explore other environments that require a high level of precision. Harb-

Wu and Krumer (2019) looked at professional biathlon athletes, as a biathlete has to perform

the exact same non-interactive task of shooting the exact same number of times. As the authors

point out, this allows exploration of within-biathlete variation (e.g., with and without being in

front of a supportive audience). Moreover, as all biathletes need to perform the precision task,

a selection effect as discussed in Dohmen (2008) is less of a problem (although individuals

may select themselves into that sports field). Their large data set covers 16 seasons including

144 World Cup events, 12 World Championships and 4 Winter Olympic Games – and the

results indicate for both genders that biathletes from the top quartile struggle more in their

home turf compared to competitions abroad.

Basketball is another environment that allows exploration of performance under

pressure in a more controlled way, especially when looking at free throw percentages as a

measure of performance (for a discussion see Cao et al. 2011). Each free throw attempt is taken

from the same location, which means that the physical difficulty is constant. In addition,

contrary to a penalty kick, the performance is not confounded by a response of another player

(e.g., goalkeeper in a penalty kick). Contrary to penalty kicks, free throws occur very frequently

and for most players. This makes it possible to explore the heterogeneity in shooting skills and

stress resistance and means that results are less driven by a selection effect. Moreover,

11
psychological factors may still matter because, as Cao et al. (2011) stress, free throws are still

a non-trivial task with failure rates for most NBA players. Their results indicate that there is

some choking but the effects (e.g., playing at home) are small. However, choking becomes

more dominant at the end of the game (decline in success by around four percentage points

when the shooter’s team is down by one or two points in the final minute). The negative effects

increase to 6.3 and 8.8 percentage points for the last 15 seconds when down by two or one

points, respectively. The choking effect is also stronger for players who are worse free throw

shooters. Toma (2017) followed up on basketball free throws looking at both females and males

at the college and professional levels. Interestingly, male college players who eventually play

at the professional level choke more in the final seconds of a close game. Toma’s argument is

that they feel more pressure to perform due to their career expectations. He also finds no

evidence of a gender difference in choking behaviour.

Cohen-Zada, Krumer, Rosenboim and Shapir (2017) explored tennis to understand

human behaviour under pressure. They looked at the effect of competitive pressure on the

likelihood of winning the game, instead focusing on the number of unforced errors or winning

shots. The strength of their measure is its objectivity:

While two observers can debate whether or not a certain shot should be considered as a forced

or unforced error, and whether or not the previous shot led to the forced error, in our case

winning a game is an undeniable fact” (p. 178).

Interestingly, they find that increasing the level of stakes reduces performance for men. They

seemed to choke under competitive pressure, while women choke less. Thus, women show

superiority in this setting regarding competitive pressure. They argue that those results are

“consistent with evidence in the biological literature that levels of cortisol, which is known to

impede the performance of both men and women, commonly escalate more substantially

among men than women in response to achievement-related challenges” (p. 188).

12
 Hickman and Metz (2015) take advantage of the ability to look at performance on the

last hole of the golf tournament on PGA tours, allowing them to observe a substantial variation

in key pressure situations. Obviously, making or missing a putt can have a considerable

influence on the finishing position and the monetary reward. Their innovation is to have this

direct link to a monetary reward, which is less in the case if you fail in your free throw

performance in the NBA. An interesting addition is that they have data on the exact location

of players’ golf balls before and after each shot (down to an inch). Their results indicate that

increasing the value of a putt by around $50,000 decreases the likelihood of a player making

the putt by one percentage point. However, that magnitude is greater for specific shots such as

those taken from five to 10 feet away. Not surprisingly, less experienced players are more

negatively affected by pressure.

One of the major issues limiting the empirical analysis of the stress/performance

relationship beyond the sports environment is measurement (measuring an individual’s

performance and then comparing it to another). Performance is not fully comparable in most

workplaces, even between two individuals doing the same job, because regardless of the metric

used to measure performance it needs to be analyzed using the same underlying characteristic

or the statistical inference erodes (Allison 1999). For example, the environmental conditions,

incentive structures, support systems, or any number of other exogenous factors may not be

identical for both individuals. Given the broad availability of information on athletes and the

relatively controlled conditions under which they compete, the use of the sporting environment

has been fruitful. However, the modern era of Big Data may make this an even better

experimental laboratory – many players now wear IoT (internet of things) devices during

competition that track heart rates, blood pressure, speed, distance, and a range of other

geotagged data that could be coupled to the on-field behavior and choices made by players.

13
This combination of new data could combine biological and physiological aspects with

decision sciences (for a discussion, see Torgler 2020).

Beyond Physical: eSports

When it comes to the human body’s non-sympathetic responses to stress, our brains cannot tell

the difference between reality and imagination. Our bodies release stimulants like adrenalin

and cortisol when we are stressed or when we imagine being stressed (Hamilton 2018). This

psychological effect also works in a number of other situations, e.g., the placebo. Under the

placebo effect we believe we have received a drug that will have a certain effect. Since our

mind is unable to distinguish between reality and imagination, it will often supply the required

sensation or, where possible, the stimulation to replicate what was expected – such as the

diminishing of pain. Studies have also shown that we can imagine our bodies fighting cancer

(Eremin et al. 2009) or recovering body function after a stroke (Kho et al. 2014) and our mind

rewires itself to make it happen. Many of the aspects of sports explored by behavioural

economists in the past have more than likely been confounded with physical aspects or skill of

the athletes being studied – one such example is sporting momentum (see e.g., Gauriot and

Page 2018, Cohen-Zada, Krumer and Shtudiner 2017). Thus, eSports offers the opportunity to

disentangle such mental and physical aspects. Psychological momentum (which will be

discussed in more detail later) can work as positive and negative momentum and alters

behaviour and performance by the perception of an individual (Iso-Ahola and Dotson 2016). It

has been argued that positive psychological momentum improves the individual’s confidence,

and by extension their competence which allows them to successfully complete the task at hand

in a better and faster manner – this then increases the individual’s expectations of success on

the next task. The problem with this argument is that it is directly related to physical attributes

of the individual, i.e., the assumption is that if the individual can complete a task faster (or

14
better) they were either not exerting maximum effort or were doing so inefficiently. The

physicality issue poses a problem when trying to compare momentum between athletes as it is

unclear if individuals were actively conserving energy for future effort (or if they were already

at maximum effort), and how we can determine the comparative differences in effort and

energy between players. Again, we could turn to eSports where physical aspects are less

important in comparison to the mental (or psychological) ones – removing one of the confounds

and making it clearer to analyse momentum.

Concepts

Prospect Theory

There are many contentious or unclear theoretical concepts that are difficult to empirically

prove, and several experiments and behavioural concepts are traditionally explored in

laboratory settings that may be better served if we can utilise the quasi-natural field aspect of

sports. For example, Prospect Theory has only occasionally been tested in the real world with

real-world losses (Page at al. 2014), but it may be much easier to test in an environment with

clear incentives, strategic actions, and real-world gains and losses. Prospect Theory relies on a

lack of adaption (habituation) of losses in terms of the individuals’ (or perhaps the groups’)

reference point (Kahneman and Tversky 1979). This point is supported anecdotally by

gamblers who often “chase their losses” or take on additional risk when they are behind but

may also take on more risk due to an endowment (house money) effect where a newly acquired

gain has not been habituated and can be gambled without fear of loss (Thaler and Johnson,

1990). However, it may be possible to explore this effect within the sporting context with

variations in risk behaviour in game (during) and out of game (after) – where post game we

would expect the players to now be in a ‘cold’ state and able to habituate the losses, but during

a game the players in a ‘hot state are likely to chase losses (see Loewenstein 2005). There may

15
also be many potential avenues from which to approach this topic, from both the individual

and team levels. Individual sports could be used to provide insight into how players habituate

losses during highly incentivised tournament matches and may be compared to choices made

during non-competitive or low incentive matches. Alternatively, this could be explored in

conjunction with the group assessment of risk (see below) with team sports – where individuals

and teams could be examined for changes in risk attitudes (behaviour) during matches.

Additionally, it might be possible to explore differences between one-shot, repeated, and

knock-out games – to see if strategic elements make a difference in choices being made.

Goal Setting

The concept of goal setting, where a fixed goal provides an aspirational point of reference, is

an extension to Prospect Theory, where the reference point moves after gains or losses have

been habituated (see Locke 1968, Locke et al. 1990). If the theory holds, then individuals are

more likely to work harder to achieve a difficult goal than they would if no goal or an easily

achieved one was set. An extension of this theory could be that during sports competitions or

tournaments (rather than single matches), losses could inspire players to strive harder or take

on more risky options to achieve the stated goal. For example, as the probability of success

wanes (decreases) the players may be willing to adopt riskier and riskier gambles in an attempt

to overcome the lower probability. This often occurs in tournaments where teams or individuals

hold an expectation of where they should finish (regardless of if that is first place, fifth, or

merely making it out of the first stages).

Interdependent Preferences

Another theory that has proven difficult to empirically explore has been that of interdependent

preferences (Pollak 1976), where the preferences of an individual are co-dependent on the

16
preferences of others. While this is fundamentally at odds with the rational choice self-

interested homo-economicus model of a utility maximiser, behavioural economics has

repeatedly shown that individuals regularly deviate from the behaviour expected from this

model. However, the sporting environment may provide us with an interesting environment to

explore what happens when both self and collective interests are present – such as wanting to

maximise their own payoff, but where others are required to reach this goal. For example,

player contracts or wages in individual sports are purely dependant on the players own

performances and skills. However, in team sports we observe that an individual’s ability to

excel in their own position is directly dependent on those around them, as such every player’s

success depends on that of another (Frank 1984). Sports data provides an unusually large

amount of interconnected data on performance and earnings that can be exploited to explore

how the contract negotiations of similar players are linked to the performance of teammates

and other players of similar skills. As such, even superstars with the highest ranking may be

concerned about the ranking and incentives of those around them (Postlewaite,1998), and not

just their absolute position or earnings. A great example of this relationship could be seen

between Michael Jordan arguably the greatest basketball player of all time and his long-time

teammate Scottie Pippen:

He helped me so much in the way I approached the game, in the way I played the game. Whenever

they speak Michael Jordan, they should speak Scottie Pippen. Everybody says I won all these

championships. But I didn’t win without Scottie Pippen. That’s why I consider him my greatest

teammate of all-time (Michael Jordan, 2020: The Last Dance).

Strategic Interactions

The analysis of sports in the sports economics methodology has been doubly focused on many

traditional and mainstream sports but there may be some significant advantages to exploring

more strategic games such as chess, go, or poker to analyse strategic and k-level thinking.

17
Strategic thinking may also open the door to explore free riding, minimax, or maximin as

sporting strategies.

(Group) Risk

Early experiments in social psychology (Stoner 1961, Bem et al. 1965) show that groups are

more likely to take on more risk than the individuals that make up those groups. Termed as a

“risky shift”, the analysed groups almost always took on more risk. It may be that individuals

feel more comfortable adopting more risk if they can share the responsibility (blame) for a

failure amongst the group, but if that responsibility falls on the individual, they are more risk

adverse. This may also be related to the concept of “group think” (Janis 1972), where the

individual may be more concerned about what others think as opposed to their own self-

assessment. Rather than question those in charge or above them on the hierarchical chain,

individuals will suppress their own views if they do not align and converge on a risk assessment

based on a focal point rather than a distribution of the group (see, e.g., Bénabou 2013). We

have observed the disasters that can occur when group think is allowed to flourish in the

financial markets. The Global Financial Crisis (GFC) was rooted in the belief that the US

housing market could not fail – until it did. This is not the first time the finance industry came

close to collapse caused by a group think situation, this problem was described as data-think,

where everyone is relying on the same data: the fact that “everyone will be wrong about the

same thing at the same time brought hedge fund Long-Term Capital Management close to

collapse in 1998” (Hill 2018)1. Thus far economics (especially behavioural) has predominately

focused on individual decision makers and how risk preference, risk type, risk seeking/aversion

has impacted their choices. Obviously, this becomes more complicated if we wish to explore

1
Hill, A. (2018). Why groupthink never went away. Financial Times, May 7th, 2018. Available from
https://www.ft.com/content/297ffe7c-4ee4-11e8-9471-a083af05aea7

18
group behaviour and attitudes, however, this may be where the sporting arena may be helpful.

Goff and Tollison (1990, pp. 6–7) define it thus: “Sports events take place in a controlled

environment and generate outcomes that come very close to holding ‘other things equal’”.

Thus, sporting events can be seen as ‘quasi-natural field experiments’ where subjects are acting

in the natural environment instead of an artificial laboratory environment (natural incentives to

perform) and players compete in an actual high-stakes contest with real incentives to be

successful (Goff and Tollison 1990). In such an environment it may be possible to explore

whether group assessments of risk differ from that of the individual. For example, an individual

may be personally risk seeking, but they play in a team that has a demonstrably risk averse to

attitude.

(Mismatched) Incentives

We know that there have been mismatched incentives in sports due to the winner take all

market (Frank and Cook 1995), where a small variation in performance or skill results in very

large changes in payoffs. To quote Ricky Bobby2 “if you’re not first, you’re last”. Effectively

summing up the problem: you are either the winner or you are nothing, and in such an

environment any advantage can make the difference. This becomes especially apparent in the

use of performance enhancing drugs (PED), extreme training regimes, and the adverse health

outcomes athletes may be willing to accept (Humphreys and Ruseski 2011). The short-term

payoffs for cheating or abusing their health must outweigh any potential long-term damage or

health concerns. The question for us to consider: is there a behavioural approach on how we

could realign incentives and payoffs? This seems to fit within a temporal discounting problem

2
Talladega Nights: The Ballad of Ricky Bobby – starring Will Ferrell and Sacha Baron Cohen, directed by
Adam McKay is a parody on NASCAR Racing (2006).

19
as current (or short term) rewards are valued much higher than future (long term) losses, where

current consumption (utility) is overweighting the future negatives.

Biases

We will next discuss a set of biases and how sports have helped to explore such biases

empirically. The list is far from complete; however, a nice overview of biases in general is

provided by Dobelli (2013). Future studies could map them in more detail with available

evidence in sports or how those biases can be analysed in detail in the sports setting. We focus

on an interesting set that shows the power of sports data in exploring commonly discussed

biases in behavioural economics. That also means that we do not intend to provide a detailed

literature overview of papers within the area that we are discussing. Such an attempt would go

beyond the aim and scope of this book chapter.

Sunk Cost Fallacy or Escalation Effects

One can only assume that sports clubs are subject to sunk cost fallacy when investing in players.

This means that they deviate from the classical economics approach, which would assume that

club decision makers would only consider incremental costs in their decision making. The

empirical design of using sports data has the advantage of holding the industry and competitive

conditions constant, and of observing the actual reactions of the decision makers (Pedace and

Smith 2013). One of the first studies to explore sunk costs in the sports environment was

conducted by Staw and Hoang (1995). They were inspired by experimental studies that

explored sunk-costs effects focusing on resource utilization (see Arkes and Blumer 1985).

When analysing the psychology of sunk costs, Arkes and Blumer (1985) find strong support

that sunk costs are robust judgement errors (e.g., psychological justification of such a

maladaptive behaviour is due to the desire to not appear wasteful (p. 125)). The innovation

20
from Staw and Hoang (1995) was exploring the sunk-cost effect in a natural organizational

setting using sports data (NBA data). They focused on the idea that people may perceive an

association between draft order and the prospect of future performance (strong expectation of

performance that persists long after the decline in court skills):

In our view, the presence of cognitive bias, commitment, wastefulness, and justification may

all be interwoven in natural situations. In the case of the NBA, taking a player high in the draft

usually involves some extremely high, often biased, estimates of the person's skills. The draft

also involves a very visible public commitment, one that symbolizes the linkage of a team's

future with the fortunes of a particular player. Moreover, the selection of a player high in the

draft signals to others that a major investment is being made, one that is not to be wasted. If the

draft choice fails to perform as expected, team management can expect a barrage of criticism.

Having to face hostile sports commentators as well as a doubting public may easily lead to

efforts to defend or justify the choice. In the end, team management may convince itself that

the highly drafted player just needs additional time to become successful, making increased

investments of playing time to avoid wasting the draft choice (p. 492).

The argument is that once the actual performance data in the NBA are available such a signal

should not provide any further information about a player’s ability (Borland, Lee, and

Macdonald 2011). Focusing on playing time, survival in the league, and the likelihood of being

traded, they found evidence for such a sunk-cost effect. For those better drafted players, they

observed more playing time, a longer NBA career, and a lower probability to be traded

(controlling for other important predictors such as performance, injury, or trade status).

Camerer and Weber (1999) extended on that study by collecting a new sample and testing

alternative rational explanations. They found an effect which was around half as strong in

magnitude and statistical strength but that supported Staw and Hoang (1995)’s basic

conclusions while improving the methodological approach via accounting for alternative

explanations. Such results are additionally supported by Coates and Oguntimein (2010) and

21
Groothuis and Hill (2004), who found that the draft number affects NBA career duration even

after controlling for performance measures. Leeds, Leeds, and Motomura (2015) further extend

on those studies by focusing on the transition between states (lottery versus nonlottery or first

versus second round picks) and allowing them to apply a regression discontinuity approach in

the hope of handling omitted variable biases or causality issues. Their regression discontinuity

results indicate that a lottery pick or first-round draft choices receive no more playing time for

their draft status (over those drafted later).

On the other hand, further evidence for a sunk costs effect has been found when looking

at Major League Baseball managers (Predace and Smith 2013) and the Australian Football

League (AFL) (Borland et al. 2011), although the AFL study only found limited evidence that

was largely concentrated around players’ initial seasons at a club. Predace and Smith (2013)

were motivated to understand whether new manager retention decisions were affected by

whether or not the poor choice was made by a previous manager. Their results indeed indicate

that poor performing players were significantly more likely to be divested by new managers

than they were by continuing managers. Keefer (2015) criticized previous studies by stressing

that players’ draft numbers are a measure of expected productivity. Keefer (2017) also

criticizes that the first-round players are not always the first player chosen by their team. Keefer

(2017) uses the NFL draft as a natural experiment, using fuzzy RD as first round players

selected near the round cut-off receive a very large wage premium (first round wage premium

of around 36 to 38%). Players selected near the round cut-offs are therefore essentially

randomized and can be used to identify a sunk cost effect via the effect of compensation on,

for example, the number of games started. One may also argue the superstars are easier to see

as their overall level of skill stands out, but as you proceed back into the general draft pool

athletes become more clustered and the skill levels between players becomes smaller which

means that it becomes much more difficult to observe skill differences. At this point the

22
decision becomes much more random – this may not actually be a reflection of players’ actual

overall potential but just on what is observably different at this early point in their career. The

results indicate sunk cost effect (a 10% increase in compensation was linked with 2.7 additional

games started). However, Keefer (2015) criticized that such a result may just represent heuristic

thinking and therefore the question emerges whether or not the sunk-cost fallacy persists when

teams receive performance feedback. He therefore explored the exogeneous variation in

compensation when players become eligible for free agency or change teams. Here, he also

finds a substantial sunk cost effect.

It is interesting to look beyond these results to other sports fields, such as soccer. Soccer

is fascinating, due to having fewer restrictions in maintaining a competitive balance between

teams in the form of salary caps or a draft system. However, using German Bundesliga data,

Hackinger (2019) finds that playing time is mainly driven by previous or predicted

performance, which means that coaches and managers ignore high transfer fees in that decision

process. Alternatively, the Oakland Athletics (A’s) baseball club fiscal and statistical strategy

was made famous by the 2011 movie Moneyball starring Brad Pitt as club manager Billie

Beane (see also Lewis 2004). The A’s have built a highly competitive and successful team,

able to compete with baseball heavyweights like the New York Yankees or the Boston Red

Sox by going against the mainstream by embracing sunk cost thinking. In recent years they

have been trading away ‘superstars’ to build a better team – in a statement to the Wall Street

Journal3 Oakland GM David Forst said:

We’ve always had to operate a little differently than everybody else. We’re never afraid to be

wrong, and if that involves trading away good players, then we’re OK with that, because,

ultimately, we have a lot of conviction in the players we’re getting in our half of the deal.

3
Story sourced from the Wall Street Journal at https://www.wsj.com/articles/the-moneyball-as-find-a-new-
inefficiency-other-teams-players-11566322019.

23
Action Bias

Behavioural economics has challenged the classical assumption that how an outcome comes

about should not matter. We often feel the emotion of regret when looking back on what turned

out to be bad decisions. There is substantial available evidence that indicates people regret

actions more if the outcome is reached by action rather than inaction (for an overview see

Zeelenberg et al. 2002). But Zeelenberg et al. (2002, p. 314) also stress the importance of

accounting for decisions made in response to things that happened earlier. This means that if

the prior outcomes were negative, you are more likely to be inclined to improve future

outcomes and therefore regret inactions more. Using scenarios from the sports domain (soccer

coach decisions), they confront subjects in an experiment with a situation in which soccer

coaches either won or lost a match prior to the current one. Their findings indicate that previous

negative outcomes provide a reason to act and that decisions not to act, which are followed by

a negative outcome, trigger regret. Thus, when a prior game was lost, a coach who acted would

feel less regret than one who did not act. At least the active coach tried to prevent (further)

losses.

In general, experiments are useful in that context as you can explore what happens if

two individuals arrive at the same negative outcome with and without acting. Such laboratory

experiments can be extended by applying physiological measures such as heart rate variability

monitors, that have been used as physiological markers of emotions (see, Dulleck et al. 2011,

2014, 2016, Torgler 2019, Macintyre et al. 2021). Sports data is also useful as it allows going

beyond using hypothetical questions which was the core methodological approach to explore

action biases (Bar-Eli et al. 2007). Bar-Eli et al. (2007) therefore looked at elite goalkeepers

during penalty kicks to explore whether an action bias existed in the real-world. Goalkeepers

are experienced in their decision-making domain and highly motivated to perform well. Penalty

kicks are also interesting in that context; due to the almost simultaneous-move game

24
characteristic, goalkeepers cannot de facto afford to wait to see how the player kicks the penalty

(Bar-Eli et al. 2007). The authors stress that the norm for goalkeepers in penalty kicks is to act,

which means jumping to the right or the left. Data from 286 penalty kicks shows that

goalkeepers chose to jump to their right or left in 93.7% of cases, while the utility-maximising

behaviour would have been to stay in the goal’s centre if the decision were based on the

probability of stopping the ball. The authors show that if a goalkeeper behaves according to

the probability matching principle, they should stay in the centre for around 28.7% of the kicks

but they only choose to stay in 6.3% of the time (see Figure 1). They therefore conclude that

action bias could explain such a behaviour: “[I]f the goalkeeper stays in the centre and a goal

is scored, it looks as if he did not do anything to stop the ball” (p. 614).

Figure 1: Penalty Jump and Kick Direction and Stopping Based on Bar-Eli et al. 2007

Jump direction
Left Centre Right Total
18.90% 0.30% 12.90% 32.20%
Left 29.6% 0.0% 0.0% 17.4%
14.30% 3.50% 10.80% 28.70%
Kick direction Centre 9.8% 60% 3.2% 13.4%
16.10% 2.40% 20.60% 39.20%
Right 0.0% 0.0% 25.4% 13.4%
49.30% 6.30% 44.40% 39.20%
Total 14.2% 33.3% 12.6% 14.7%
Source: Bar-Eli et al. 2007, pp. 612-613. Notes: The chances of stopping a
penalty (number of balls stopped/total number of jumps to that direction)
are in green, jump direction is in black.

Future studies could explore whether the level of action bias is influenced by specific

contextual factors (e.g., importance of the game) or individual characteristics (e.g., experience,

gender, age, etc.). Bar-Eli et al. (2007) also notes issues around the dynamics: “If goalkeepers

will always choose to stay in the center, however, kickers will start aiming all balls to the sides,

and it will no longer be optimal for the goalkeeper to stay in the center” (p. 616). This may also

25
be linked to the general question whether biases attenuate, fully disappear, or even reverse once

they are reported to a broader audience (see, e.g., Schwert 2003).

Outcome Bias

A commonly discussed bias in behavioural economics is outcome bias, which means that we

evaluate decisions based on the results rather than the actual decision process itself. Sports are

an interesting setting to explore outcome biases because of the importance of performance

appraisal. Sports data allows going beyond the lab, which has been the dominant method to

explore outcome biases and performance appraisal. Kausel et al. (2019) focused on penalty

shoot-outs as the outcome of a penalty shoot-out (who wins or loses), which seemed to be

unrelated to actual in-game performance beforehand (performance at the individual player

level as well as the team level). The ability to work with these data is a significant improvement

over working with field data, as in the latter we are faced with the problem that actual

performance is correlated with outcomes, which makes it hard to explore an outcome bias.

Such independence between performance and outcome can then be linked to football players’

subjective performance ratings. Kausel et al. (2019) therefore tested whether players on the

winning team received better ratings than those on a losing team. Their data (which were

derived from major soccer tournaments such as FIFA World Cup, UEF European

Championship, and UEFA Champions League) indicate that winning the penalty had a positive

effect on reporters’ performance ratings, even after excluding players who participated in the

penalty shoot-out. Beyond that, they found such an effect remains when using fixed effects

(within-players’ design).

Expert and Judge’s Biases

26
Merkel et al. (2021) explored performance appraisal by looking for signs of optimism and/or

positivity bias in rating semi-annual performance appraisals in the youth academy of a German

Bundesliga. The interesting thing is that this allowed exploration of three types of evaluations:

a rating of predicted future performance, a rating of remembered performance during the last

half-year, and record of instantly reported ratings of the actual performance in individual

matches. Those ratings are important as they affect athletes’ possibility to progress to the next

age group. Their results indicate that predicted and remembered performance ratings

significantly exceeded actual ratings. Such a deviation is more pronounced for the predicted

performance, which indicates some asymmetry between looking forward and backward. Such

biases may be unintentional, while other biases can be more intentional.

Nationalistic biases have offered an interesting avenue for past explorations. In general,

events such as the Olympics provide a rich data source to explore such biases; as various fields

such as ski jumping, figure skating, diving, etc. rely on judges’ scores. As Zitzewitz (2006, p.

68) points out:

[I]n most settings, attempts to study favoritism empirically would be frustrated by the difficulty

of observing where one should expect favoritism (e.g., who is “friends” with whom).

The sports environment, on the other hand, allows us to explore whether judges are

nationalistically biased. Beyond that, Zitzewitz (2006) explored whether such a bias varies with

strategic considerations, which would indicate intentionality. The sports setting also provides

the opportunity to explore incentives based on different institutional conditions. Zitzewitz

(2006, p. 69) discusses the differences between ski jumping and figure skating. In ski jumping,

judges are chosen by the Federation International du Ski (FIS), while national federations

choose the judges to be represented at Olympics in figure skating. Zitzewitz (2006) even

observes that in figure skating, bloc judging or vote trading is found. One would therefore

expect more nationalistic biases in figure skating. In addition, as FIS selects judges based on

27
pre-Olympic events, one may expect to observe more nationalistic biases in the Olympics

compared to the pre-Olympic events (which he finds to be true). In general, Zitzewitz (2006)

finds a relatively large effect. For example, a nationalistic bias in figure skating translates to

an average of 0.7 higher ranking position placement. He stresses that some of his findings make

it difficult to rationalize with tastes (e.g., for a particular national style of skating) or

unconscious biases:

Examples include the fact that the national identity and past judging bias record of the other

panel members appears to affect scores or the fact that biases vary in a way that accords with

judges’ career concerns (p. 70).

Emerson, Seltzer and Lin (2009) looked at diving competitions from the 2000 Summer

Olympic Games to see whether judges have preferences for individual divers due to, for

example, style. They stress that a residual from their model represents the difference between

judges’ scores and the predicted score of judges for a diver given the nationalities of the judge

and the diver. They conduct an analysis of variance predicting the residuals using individual

divers as explanatory factor, which means that the coefficient would indicate a judge’s

preference for individual divers due to unobserved reasons after controlling for nationalistic

preferences (p. 130). Their results indicate that only one judge reported differences that are of

statistical significance. Overall, they observe strong evidence of nationalistic favouritism.

Further studies judges’ or referees’ biases looked at gymnastic judges (Flessas et al.

2015, Heiniger and Mercier 2019), combat sports (Myers et al. 2006), soccer (Torgler 2004,

Pope and Pope 2015). Ansorge and Scheer (1988) looked at gymnastic competition at the 1984

Olympic Games and found judges not only overscore gymnasts from their own countries, but

also underscore gymnasts from countries who are in close competition with their own country.

Campbell and Galbraith (1996) find some evidence that the bias against Olympic figure-skaters

is stronger for medal contenders than for competitors who are less strong.

28
 Zitzewitz (2014) also explored the interesting policy adjustment of reducing

transparency among the International Skating Union (ISU) via no longer reporting which judge

gave which score after the vote trading scandals in the 1998 and 2002 Olympics. ISU hoped to

reduce outside pressure on judges in order to reduce favouritism and corruption, but Zitzewitz

(2014) was able to show that nationalistic bias and vote trading actually increased slightly

without being statistically significant after the reforms. He points out that “[i]f nationalistic

bias has increased in importance relative to vote trading, we might expect to see a single

positive outlier score when a compatriot is on the panel” (p. 23).

The literature on referees’ home court biases is closely connected to this literature,

particularly in the area of soccer. The paper by Garicano, Palacios-Huerta, and Prendergast

(2005) on extra allowance time when the home team is behind by one goal (compared to when

being ahead by one goal) influenced many studies. Dohmen and Sauermann (2016) provide an

excellent overview of that area of research. When looking at extra time they conclude that there

is evidence for systematic referee bias in the second half of the game. The effect is strongest

when the home team is one goal behind before the stoppage time begins. In addition, the bias

is larger in Spain compared to Germany, possibly due to the higher travel distance (longer than

700 kilometres, compared to 400 in Germany). Dohmen and Sauermann (2016) also review

other decisions such as awarding goals or penalty kicks, as they have a more immediate impact

on the outcome of a game. The overview indicates that home-biases are visible for penalty

kicks. Home teams benefit from a larger fraction of awarded penalty kicks that are wrongly

given or disputable when being behind by one goal (see Dohmen 2008b). Similarly, various

studies (but not all) found a bias regarding the award of yellow and red cards, controlling for

players’ behaviour, noting that bias is triggered by the crowd density (Dohmen and Sauermann

2016). Overall, Dohmen and Sauermann (2016) summarize the biases into two categories:

those driven by social and those by material payoffs. Social payoffs are linked to the size and

29
composition of the supporting crowd, or distance to the crowd. However, they stress that social

forces can be partly offset by material payoffs, such as increasing the referees’ wages or by

better monitoring their decisions.

The question of what happens with referee bias once the stadium is empty is an

interesting avenue to consider. After a serious act of hooligan violence between supporters

from Calcio Catania and Palermo Calcio in 2007, the Italian government forced teams to play

their home games without spectators if they had stadiums with deficient safety standards.

Pettersson-Lidbom and Priks (2010) took advantage of the situation to study the difference in

an empty stadium. Using data from Serie A and Serie B for the season 2006/2007 up to the

point when all teams (apart from Catania) played in front of spectators again (842 games), they

find that referees significantly change their behaviour in games played without spectators.

Home teams are punished less harshly than the away teams with spectators, but more harshly

without spectators. Similarly, Bryson et al. (2021) take advantage of the COVID-19 pandemic

as a natural experiment which induced a near-complete absence of fans in sporting arenas.

Using a large data set from 6481 football games and 17 leagues played before and after the

mid-season shutdown, they find that the absence of crowds reduces home advantages.

Significantly fewer yellow cards were awarded to the away team without a crowd, narrowing

down the gap between the home and away teams by around a third.

Hot Hand Fallacy and Momentum Effect

A natural avenue in the history of behavioural economics was the exploration of whether we

see patterns where there is actual randomness. Our brains are usually well-equipped to see

patterns, as such a skill increases survival chances. However, one cannot exclude the possibility

that we are subject to cognitive illusions. In one of the most famous sports papers, Gilovich,

Vallone, and Tversky (1985) investigated whether or not there is a hot hand in basketball by

30
looking at the Philadelphia 76ers in the 1980-81 season, as a large number of individuals

(including sports experts) believe that a player has a better chance of making a successful shot

after having made the last two or three. The innovative approach looked at conditional

probabilities for nine players (shooting percentage of having missed or hit the last shot, last

two, and last three). Their study could not find a hot hand. In fact, eight of the nine players’

probability of a hit was actually lower following a hit than a miss.

That result inspired a very large set of different studies, which are too numerous to

properly discuss. One goal of the studies was to achieve more controllability. The problem is

that the game itself is subject to a rich context where other effects may take place. For example,

a “hot shooter” may believe they are on fire and may therefore take more difficult shots, which

may reduce the success rate. Or the opposing team may start guarding a “hot shooter” more

closely, reducing future success rates (Koehler and Conley 2003). Consequently, studies

focused on measuring and analysing free throw successes. However, this focus is not free of

problems. First, there is a high probability of success in free throws (around 75%) and this

brings the problem of potential time lags (Koehler and Conley 2003). New free throw

opportunities are often too far apart. Thus, researchers started looking at long distance shootout

contests such as the annual NBA competition where eight of the best 3-point shooters compete

against each other, taking five uncontested shots for five pre-determined sports around the 3-

point arc, allowing for sixty seconds to finalize all the shots. A hot hand hypothesis would

suggest that a shooter would have fewer runs (e.g., HHHHH (one run) versus HMHMH (five

runs) but that does not seem to be the case (Koehler and Conley 2003). When comparing

expected and actual runs, Koehler and Conley (2003) indicate a lack of a hot hand effect on

NBA Long Distance Shootouts. However, Burns (2004) emphasizes that the hot hand

phenomenon has two separate components: the hot hand belief regarding dependence and a hot

31
hand behaviour following streaks. In other words, invalidating a belief does not necessarily

invalidate the behaviour that is based on that belief (p. 300):

There is no doubt that the beliefs people hold play an important role in their decision making

and thus identifying those beliefs is useful. However, unless one thinks it is better to score less

in basketball than would be possible if a simple cue was given some weight, then it appears that

a research focus on belief without regard to behavior has led to the mis-analysis of an important

decision-making phenomenon(p. 327).

Stressing the advantage of analysing reason in Gigerenzer’s (2000) framework of adaptiveness,

Burns (2004) uses simulations to show that streaks are valid allocation cues for deciding who

should be given a shot, allowing the team to score more.

Studies around the hot hand fallacy or a momentum effect have explored beyond

basketball (namely in baseball, tennis, golf (including golf putting), soccer, volleyball, darts,

tenpin bowling, or horseshoe tossing), but a key problem remains that such studies fail to

understand the actual cognitive processes around it in more detail (Alter and Oppenheimer

2006). Focusing on neuroscientific insights or identifying settings are interesting avenues that

have a biological background. Burns (2004), for example, cites neuroscientific evidence that:

demonstrated that different areas of the brain are more activated by streaks than by nonstreaks.

Not only do specific areas of the brain react to streaks, but the strength of the signal is related

to the length of the streak. If the brain is wired to notice streaks, then it is unsurprising if it is

also found that people utilize streaks in making choices. Furthermore it also implies that doing

so is probably useful in some way (p. 299).

Page and Coates (2017) used professional tennis matches to understand the importance of a

winner effect that could be driven by testosterone changes, arguing that the winner effect might

be mediated by a physiological feedback loop: “winning leads to higher levels of, or increased

sensitivity to, testosterone, which in turn raises the likelihood of further victories” (p. 531).

However, testosterone should not drive the results for female tennis players (females have 10

32
to 20 percent of the testosterone levels of males, Coates 2012). Their results indeed indicate

sex differences. Men who won a closely fought tie-break had around 60% chance of winning

the following set, while this winner effect did not exist among women. The importance of

focusing on the biology has been advocated by Coates (2012), looking in detail at traders. He

stresses that:

economics needs to put the body back into the economy. Rather than assuming rationality and

an efficient market – the unfortunate upshot of which has been a trading community gone feral

– we should study the behavior of actual traders and investors, much as the behavioural

economists do, only we should include in that study the influence of their biology (p. 36).

In general, behavioural economics has failed to fully explore the possibilities of biology

(Torgler 2016). Several recent contributions have shown that we can learn a lot studying

humans from an evolutionary perspective (see, e.g., Wilson 2019). Surprisingly, several

behavioural economists have been very critical regarding evolutionary psychology. Thaler

(2015), for example, stresses in his book Misbehaving that “accepting the theory of evolution

as true does not mean that it needs to feature prominently in an economic analysis. We know

people are loss averse; we don’t need to know whether it has an evolutionary explanation” (p.

261). Tversky is also said to have pointed out: “Listen to evolutionary psychologists long

enough and you’ll stop believing in evolution” (Lewis 2017, p. 336). This somehow goes back

to the academic fight or intellectual battle between Gigerenzer and Tversky and Kahneman

(see in detail Lewis 2017). Gigerenzer emphasized the importance of adaptive theories and the

relation between the mind and the environment rather than the mind alone (for a discussion see

also Torgler 2021b). In his Psychology Review article Gigerenzer (1996) stressed that “the

issue is not whether or not, or how often, cognitive illusions disappear. For him, the focus

should be rather the construction of detailed models of cognitive processes that explain when

and why they disappear” (p. 592). Gigerenzer (2004) also recollects the following discussions

with Herb Simon:

33
 Herb applauded the demonstrations of systematic deviations from expected utility by

Kahneman, Tversky, and others. But what did he think when the followers of Kahneman and

Tversky labeled these demonstrations the study of ‘‘bounded rationality?’’ I asked him once,

and his response was ‘‘That’s rhetoric. But Kahneman and Tversky have decisively disproved

economists’ rationality model.’’ Herb was surprised to hear that I held their notion of cognitive

illusions and biases to be inconsistent with his concept of bounded rationality. I think he liked

their results so much that he tended to overlook that these experimenters accepted as normative

the very optimization theories that Herb so fought against, at least when the results were

interpreted as cognitive illusions. A true theory of bounded rationality does not rely on

optimization theories, neither as descriptions nor as norms of behavior… A systematic

deviation from an ‘‘insane’’ standard should not automatically be called a judgmental error,

should it? ‘‘I hadn’t thought about it in this way,’’ Herb replied (pp. 396-397).

(Limited) Attention or Hyper Attention

Players and fans alike are always looking for a win, and anything that might help that outcome

is warmly embraced, however, anything that ‘might’ impinge upon success quickly becomes

anathema, even if it is completely illogical. For example, the colour green is seen as a major

problem for equestrian riders and NASCAR drivers – even a hint of green being worn in the

audience spells bad luck for a horse and its rider4, because green was linked back to major

NASCAR accidents in the 1920’s5. While some of the superstitions held by players when it

comes to certain colours may seem odd, there is a good amount of research to back some of it

up. While it sounds like a myth, being in the red corner for Olympic combat sports such as

boxing, taekwondo, Greco–Roman wrestling, and freestyle wrestling, does statistically

improve the chances of winning (Hill and Barton 2005). This colour effect has been shown to

4
Sourced from Eclipse magazine https://eclipsemagazine.co.uk/5-horse-racing-superstitions-explained/
5
Sourced from https://jalopnik.com/how-the-color-green-became-a-deadly-bad-luck-superstiti-1763008917

34
carry over to eSports with ‘death matches’ in 2004’s First-Person Shooter (FPS) game Unreal

Tournament. Ilie et al. (2008) found that over a 3-month period (1347 observations) the red

teams won 54.9% of the matches – even though players were anonymous, and the player

avatars were visually identical except for the team colours. Piatti et al. (2012) undertook an

analysis of the red effect on professional Rugby League (Australia) teams over 30 years (1979-

2008), which covered 5604 individual matches. They found that wearing some amount of red

resulted in teams winning more often than teams without any red in their jersey stripes –

specifically, that by shifting from no-red to a little increased the probability of winning by 4.3

percentage points and by shifting from a little red to red being a major colour in the strip

increased the probability of winning by 7.5 percentage points. While the effect has been shown

to be present across several sports, the underlying cause of the effect is still not clear, i.e., is it

as simple as just the increased visibility of the colour, or is it biologically encoded in our DNA

from a millennium of evolution? It would be interesting to extend this research to understand

the role colour in all aspects sport.

One of the oldest sports training tips across the world is probably “keep your eyes on

the ball,” it does not matter which sport was being talked about – but one would assume it is

relevant for non-ball sports as well i.e., watch the puck (clay pigeon), focus on the target

(bullseye), etc. But what happens when athletes become stressed and take that maxim to

extremes? As discussed above, high levels of stress and pressure can lead to individuals making

sub-optimal decisions – one of the reasons for this is hyper-vigilance (Schultz 1966, Janis and

Mann 1977). Essentially, this is when athletes begin to second-guess their choices and switch

from the usual automatic and instinctual behavioural responses into a more laborious and time-

consuming step-by-step thought process (Beilock and Carr 2001, Bourne and Yaroush 2003,

Lehner et al. 1997). The hyper attention focus on every detail, rather than natural processes,

results in slowed reaction times and degraded performance, reverting to maladaptive

35
overthinking (Epstein and Katz 1992) or obsessive focus on singular aspects or tasks to the

neglect of all else. This can lead to an inability to respond to or quickly react to changes outside

the focal point – a possible example of this is Biaggio’s 1994 FIFA World Cup finals penalty

shot miss, where he was so focused on correctly striking the ball, he may have not given enough

attention to aiming where he was kicking the ball (which went flying well above the cross bar).

Challenging Topics: Some Examples

In this section, we will discuss how sports data can provide a tool for the exploration of

challenging topics, particularly if we try to go outside the laboratory or observe human

behaviour in the labour force. We will briefly discuss three examples.

Cooperation

Team sports data allow us to explore in detail how players cooperate and interact, as one is

able to see how athletes interact together and under what circumstances (e.g., decisiveness of

the game situation). While the cooperation literature is quite extensive (for an overview, see,

e.g., Christakis 2019), the dynamics of cooperation are still not well enough understood beyond

a lab or simulation setting. In repeated interactions, a higher payoff can be achieved through

conditional cooperation or reciprocity. In other words, following rules of good behaviour can

be a good strategy. Game theorists have explored this question in detail. Through repeated

interaction, you can achieve peaceful cooperation; for example, as exemplified by the Folk

theorem. Future expected punishment can enforce cooperation, despite a strong short-term

conflicting interest (incentive to cheat). In other words, “prospect of vengeful retaliation paves

the way for amicable cooperation” (Nowak and Highfield 2011, p. 29). Various punishment

strategies can then be discussed, as by Axelrod (1984), via a tournament or competition that

resulted in identifying the winning strategy. Tit-for-tat, developed by Anatol Rapoport,

36
performed best in this setting. Nowak and Highfield (2011) report fascinating simulations

conducted by Martin Nowak. His innovation was introducing chance (cooperation with a

certain probability). The most powerful strategy in the dynamic interplay of cooperation and

selfishness in a world that started with primordial chaos (random strategy) was a generous tit-

for-tat strategy. Always meet cooperation with cooperation; but when faced with defection,

cooperate for one in every three encounters. In other words, the recipe for forgiveness was

probabilistic. In another simulation the winning strategy was: “If we have both cooperated in

the last round, then I will cooperate once again. If we have both defected, then I will cooperate

(with a certain probability). If you have cooperated and I defected, then I will defect again. If

you have defected and I have cooperated, then I will defect” (p. 43).

We argue that the sports environment can also provide an interesting environment in

which to understand cooperation and free-riding behaviour. Accordingly, we discuss an

interesting study by Brouwer and Potters (2019) that focused on cyclists’ breakaways. During

such a breakaway, riders are required to cooperate if they are keen to build a lead over the

chasing peloton that has more manpower. But there is a social dilemma in this situation. As air

resistance can fatigue a rider, anyone in this newly formed “team” may try to minimize being

in the leading position to have more energy for a final sprint (effort saving strategy). A rotation

formation would be a highly cooperative formation to deal with free-riding incentives. Free-

riding would mean refusing to lead the group, or underperforming by exerting less effort when

at the front6. Such shirking is, as Brouwer and Potters (2019) found, harder to detect. The

authors find a positive effect of group size and group strength on breakaway success, but the

effect is concave (reaching the optimal level at 26 riders, meaning that adding another rider

6
There is also the strategic issue where rider from leading teams try to join breakaways with the intention of
slowing the overall pace to allow the peloton to catch up within a certain distance to the finish or to burn out the
breakaway. Alternatively, they may also seek to ensure the peloton does success if the breakaway does not contain
any GC contenders.

37
reduces the chances of success). In those situations where the benefits of free-riding are

smaller, such as during mountain stages, breakaways are more likely to be successful.

Furthermore, we have previously stressed that interactions can be clearly measured in

the sports context. This means we can explore how individual characteristics (e.g., dominance,

beauty, experience, being new in a team, etc.) are connected to collaboration.

Emotions

Understanding emotions is important for behavioural economics as it contributes to a better

understanding of human nature. As Simon (1983) emphasized, a general theory of thinking and

problem solving requires incorporating the influence of emotion. We have substantial evidence

in cognitive psychology or neuroscience that emotions influence memory, judgment, or

decision-making. They have important functionalities that helped humans survive in meeting

threats, challenges, and opportunities. They provide guidance in providing rapid and reliable

information, acting as communication mechanisms in our social interactions and therefore

acting as coordination tool (Keltner and Lerner 2010). Elster (1998) classified emotions into

social emotions (e.g., anger, hatred, guilt, shame, pride, admiration, or like), counterfactual

emotions (unrealized possibilities such as regret, rejoicing, disappointment, or elation),

anticipatory emotions (e.g., fear or hope), realized emotions (e.g., grief or joy), and material

emotions (e.g., envy, malice, indignation, or jealousy). Material emotions are a particularly

interesting area when exploring emotions with sports data. There is substantial evidence that

your relative income situation is connected to positional concerns and can affect your happiness

or wellbeing or can trigger envy and jealousy (for a discussion, see Frey et al. 2013). One

empirical challenge is identifying the proper reference group, another is the ease with which

one can explore behavioural responses in the work environment due to positional concerns.

The sports environment provides a unique opportunity to explore whether an increase in

38
income differences, which is transparently available for some disciplines such as basketball,

leads to a performance increase or decrease in a competitive environment that often encourages

social comparisons (Frey et al. 2013). Looking at basketball (NBA) and soccer (German

Bundesliga) Frey et al. (2013) find support that relative income disadvantage is correlated with

a decrease in individual performance. Such results are interesting from a policy perspective if

those consequences are also found in other work environments (e.g., how to design pay-for-

performance mechanisms to encourage performance and cooperation within teams). Schaffner

and Torgler (2008) show that closeness affects positional concerns when comparing different

reference groups using NBA data. The strongest effects of positional concerns on performance

are found among players with similar work profiles (playing the same position and being a

teammate) compared to other characteristics such geographical closeness, age, and experience

closeness.

Social capital

Social capital has widely explored how social capital creates human capital (Coleman 1998,

Paldam 2000), lubricates economic exchange (Putnam, 1983), builds networks, and creates

trust (Coleman 1988, Fukuyama 2003, Portes, 1998, Woolcock and Narayan 2000). While

sport has been examined through the lens of social capital theory, for the most part it has been

outward looking. Specifically, researchers have explored how sport is used to engage with

society to better the lives of players or the community, e.g., community development (Skinner

et al. 2008, Walseth 2007), health and wellbeing (Kawachi, et al. 2008, Kim et al. 2020),

participation rates (Kumar et al. 2018), or volunteerism (Darcy et al. 2014, Kay and Bradbury

2009). However, there seems to be very little research that looks inward to the effect of social

capital on the way that players interact with each other or use their social capital to be more

successful. Even though little research has looked internally, discussions have pointed to areas

39
that would be of great interest, for example, if social capital leads to more efficient transactions

through access to more information, it should result in coordinated activities for mutual benefit

and a reduction in the likelihood of opportunistic behaviour (Dasgupta 1999). The question

remains if this would be relevant in competitive interactions as well as those of mutual benefit?

Social capital theory has been explored in many real-world environments but is always

difficult to truly capture what is occurring, due to the amount of noise and other observable

events. However, it may be possible to explore such theories in sport, specifically in strategic

games where we could observe changes in player behaviour based on increased levels of social

capital. The additional advantage here is that the sporting environment enables all other factors

to be held constant. Poker tournaments is one sport in which we could observe players’

strategies and styles as they play against differing opponents where they have greater or fewer

interactions over time (interaction being a proxy for social capital, or at least social experience).

Some of this could be explained with information theory, where players with more information

on the competitors should result in more even or close competitions, and by extension it should

be less likely that players directly engage each other and pick hands on which to compete.

However, social capital would explain why these players engage in a form of coopetition where

rather than compete directly with each other, they collaborate with each other to remove other

players from the table – increasing the probability of winning overall.

One of the benefits of social capital in a strategic competitive environment like poker

could be that players who have built up capital between them may be able to compete at a lower

cost with each other than we would expect otherwise. For example, in multiplayer hands

players with social capital would not ‘push’ as hard at each other and may engage in a type of

strategic interaction known as coopetition or cooperative competition (Nalebuff and

Brandenburger 1997). Rather than a zero-sum outcome, players may be willing to engage in

this type of activity to lower losses even whilst competing. This theory has been very difficult

40
to explore in the real world as data where companies are engaged in coopetition are difficult to

obtain and few laboratory experimentations in economics have been published (see

management study by Kraus et al. 2018).

Back to the Future

From Chess to AlphaGo and Beyond

The miniaturization of electronics was led by the invention of the transistor in 1947 by Bell

Labs, replacing the large vacuum tubes used up until that point. One of the unexpected side

effects of this process was enabling computer technology to be used beyond its intended goals.

This included the creation of the first ‘home’ computers, and computers being used in games

like chess (Los Alamos chess in 1957). For the most part, these initial offerings were clunky

and relatively easy to beat by human players, mostly due to the lack of a single winning strategy

and the near infinite number of move options available to players. This began to change in the

late 1970s with the Bell Labs offering (Belle) regularly beating Master level players. Over time

the games have become better, mostly through brute force processing, i.e., memorising

countless winning games and the pruning of irrelevant outcomes as moves are made. More

recent systems have deviated from this blunt approach in lieu of more nuanced systems using

artificial intelligence (AI) to play more like a human and anticipate moves. The first generation

of these systems began with Deep Blue (IBM), which lost to world champion Garry Kasparov

(4-2) in 1996, but after major upgrades beat Kasparov in 1997. Interestingly, the attempt to use

AI to play chess was the easier option (and possibly based on Western bias of chess being the

difficult game): Go is the much harder game for computers to solve. Go originated in China

sometime around 1000BC and is still widely popular today in eastern Asia. It is played on a

larger board (generally 17x17 or 19x19) with a greater number of possible moves. Estimates

41
of the number of legal moves7 available on a 19x19 board is approximately 2.08x10170. This is

vastly more than chess and has the additional problem that a player can move to any open space

on the board and is not restricted to the moves available to the remaining pieces. Given the

massive numbers of potential moves it is not possible to follow the same strategies used by

early Chess programs, thus Go requires a significant degree of intuition to pre-empt the

opponent’s strategy.

Historically, chess offered an ideal lab setting to test and explore cognitive processes

(for a detailed discussion see Rasskin-Gutman 2009). In his autobiography Models of My Life,

Herb Simon (1996) stresses that:

For most of us those of us who have not won million-dollar lotteries, or suffered sudden

crippling accidents life is much like the chess game. We make hundreds of choices among the

alternative paths that lie before us and, as the result of those choices, find ourselves pursuing

particular, perhaps highly specialized, careers, married to particular spouses, and living in

particular towns. Even if we point to a single event as the "cause" of one of these outcomes,

closer scrutiny of the path we have trod would reveal prefatory or preparatory events and

choices that made the occurrence of the critical event possible (p. 113).

Pioneers of AI such as Allen Newell, Herbert Simon, and others realized that chess could be

used as a vehicle to try simulating thought processes via computer programming8. A good

example is Allen Newell’s (1954) paper The Chess Machine: An Example of Dealing with

Complex Task by Adaptation. He was interested in the problem of playing good chess as an

ultra-complicated problem which requires thinking about mechanisms and programs necessary

to handle such problems. His approach was to use a broad collection of rules of thumb (e.g.,

chess principles to follow, measurements to make, what to do next, how to interpret those rules

7
See Tromp and Farnebäck (2016) Combinatorics of Go. Working Paper p.22 (Table 4). Accessed from
https://tromp.github.io/go/gostate.pdf
8
For a discussion on pioneering work on computer chess by Claude Shannon, Alan Turing, Alex Bernstein, and
his authors, see Newborn (1975).

42
of thumb etc.) and focus on describing how such a set of rules is defined and organized to

achieve solutions: “the intent to see if in fact an organized collection of rules of thumb can pull

itself up by its bootstraps and learn to play good chess” (p. 23). Newell and Simon (1972) also

derived insights from how individuals played chess, using a protocol in which persons were

asked to talk aloud, mentioning moves considered and aspects of the situation. Out of that

problem-behaviour they generated graphs that show the person’s searches in a space of chess

positions. They were fascinated to discover what an experienced chess player was able to see

or perceive when looking at a chess position. Chase and Simon (1973) concluded in their

analysis that superior performance is achieved from the ability to encode positions into larger

perceptual chunks, which consisted of familiar sub-configuration of pieces. Thus, selective

search guided by heuristics can compensate for being subjected to bounded rationality (Simon

1996).

Adriaan de Groot’s (1965) Thought and Choice in Chess9 was a pioneering book on

the cognitive processes of chess; outlining an attempt to understand the thought processes

underlying skilled chess playing by also using thinking aloud protocols with experimental

sessions already held between 1938 and 1943:

[A] subject was presented with an unfamiliar position taken from an actual tournament or match

game and asked to find and play a move as though he were engaged in a tournament game of

his own. The verbal report was to be as full and explicit a rendering of the subject’s thoughts

as possible, to include his plans, calculations, and other considerations leading to the move

decision (p. v).

Rasskin-Gutman (2009) acknowledges that de Groot’s research helped us understand

that grand masters have a larger amount of chess knowledge that helped them to focus better

on general patterns of the position and identify particular characteristics.

9
For follow up books see, e.g., Avni (2004) or Aagard (2004).

43
 Herb Simon (1996) also stresses in his autobiography that he always celebrated

December 15, 1955 as the birthday of heuristic problem solving by computer, because it was

the moment when we knew how to demonstrate that a computer could use heuristic search

methods to find solutions to difficult problems. According to Ed Feigenbaum, who was a

graduate student in a course I was then teaching in GSIA, I reacted to this achievement by

walking into class and announcing, "Over the Christmas holiday, Al Newell and I invented a

thinking machine." We were not slow in broadcasting our success. In a letter to Adriaan de

Groot, on January 3, 1956, I reported: You will be interested to learn, I think, that Allen Newell

and I have made substantial progress on the chess-playing machine except that at the moment

it is not a chess-playing machine but a machine that searches out and discovers proofs for

theorems in symbolic logic. The reason for the temporary shift in subject matter is that we

found the human eye and the portions of the central nervous system most closely connected

with it to be doing too much of the work at the subconscious level in chess-playing, and we

found this aspect of human mental process (the perceptual) the most difficult to simulate.

Hence, we turned to a problem-solving field that is less "visual" in its content. Two weeks ago,

we hit upon a procedure that seems to do the trick, and although the details of the machine

coding are not yet worked out, there seem to be no more difficulties of a conceptual nature to

be overcome. By using a human (myself) to simulate the machine operating by rule and without

discretion this simulated machine has now discovered and worked out proofs for the first twenty

five or so theorems in Principia Mathematica. The processes it goes through would look very

human to you, and corroborate in many respects the data you obtained in your chess studies (p.

2006).

Deepmind’s10 solution was to adopt an AI strategy with the development of AlphaGo,

and its ability to beat the best human player in the world was not programmed – AlphaGo

learned how to play and win by itself. This is machine learning AI at its best: the system was

not programmed in any way to play Go. Instead, it started by watching thousands of games and

10
Deepmind is a London based AI lab that is a subsidiary of internet juggernaut Google.

44
learnt the game through observation. Thus, AlphaGo is classified as a General AI. It was not

programmed or taught to do a task; it used its own ability to learn in order to gain information

about the game. This AI revolution demonstrated that not only can machines learn to imitate

human behaviours but can also anticipate them, and because it is a General AI it can potentially

learn to do anything that it is presented with.

What happens next should also be very interesting for behaviouralists. If General AI

can learn to play any game (once we provide it with some framework) then the scope for its

interaction with humans and our response may be astounding. One of the general issues we

have had running experimental economics was the absence of realism (the abstractness) and

our inability to create realistic experimental conditions where we could elicit real human

responses and behaviours. The use of computing in experiments (over pen and paper) means

that we can observe every aspect of an experiment. By adopting General AI, we could start to

run experiments that might be otherwise impossible with real people. For example, high-cost,

high-risk, or life-and-death situations experiments that would be too dangerous or stressful for

humans to undertake could be replaced with AI participants. Additionally, we could observe

the impact of any number of changes to how sports are played using AI players (e.g., rules

changes) or learn human behaviours that are difficult for most of us to understand.

Poker AI

Poker is often thought of as simply a game of probabilities or risk, and as such it should be

easy to generate unbeatable computer players. However, much like the game of Go, it is the

human elements of the game that make it so hard to emulate human players. One of the central

skills of the game is bluffing (misleading opponents), which is not only about the players’

ability to control tells (ticks that give away the bluff) but also the correct betting sequences. As

early as 1998, Billings et al. (1998) flagged poker as a potential testbed for AI describing it as:

45
 a game of imperfect knowledge, where multiple competing agents must deal with risk

management, agent modelling, unreliable information and deception, much like decision-

making applications in the real world (p. 228).

This prediction become a reality when the poker program dubbed ‘Pluribus’ learned to play

six-player no-limit Texas Hold’em and performed significantly better than humans over the

course of 10,000 hands of poker (Brown and Sandholm 2019). This demonstrated that a

General AI is able to learn just about everything we can think of throwing at it; thus, the

question for us is, what do we want to do with it in the future?

AI and eSport

While the idea of eSport is relatively new to most people, it has been around in some form

since at least 1972, through a Space Invaders tournament with over 10,000 contestants

(Picknell 2019)11. eSports are one of the few sports not based on physical attributes such as

strength, but on reflex speed and mental acuity. This means that size and strength are no longer

factors, and everyone, regardless of age or gender, can compete on an even playing field. Even

so, there still is some inherent bias as computer games have long been seen as something that

boys do rather than girls, resulting in a much higher pool of male players versus female on a

competitive level. There also appears to be a fundamental difference in the type of games

preferred by male and female players. While the economics of the business (prize money,

sponsorships, viewership, etc.) have been explored, the players and their decision making have

remained relatively untouched. eSport represents an untapped field for analysis and potentially

the running of experiments – as these computer-based environments lend themselves to the

development of specific types of choices and behaviours made by players. That is, modules

11
Accessed from https://learn.g2.com/esports

46
and game designs could be specifically set up to test decision-making, collaboration, altruism,

and a broad range of other human insights.

Another major advantage of eSport tournaments is that they are played live or live

streamed players are seen, while the online version of the games are played anonymously. This

results in a comparable environmental difference; because in one the player is an anonymous

avatar, and in the other the player is an identifiable person. This poses some interesting

questions about how decision-making, and behaviour could be impacted by information and

identity, i.e., do players change behaviour if all they know is an avatar? Some previous research

has indicated that players endow their avatars with their own beliefs and social norms, but there

is also a likelihood that some players use their avatar as an aspirational representation of

themselves (Praetorius and Görlich 2020, Ratan 2020, Wiederhold 2013). This could include

choosing avatars of a different gender or race or choosing to behave in a matter that they would

like to in the real world but feel for whatever reason they are unable. Running experiments in

such an environment would enable participants to be much more invested in the outcomes and

may elicit a more realistic response than we would observe in purely laboratory experiments.

Another advantage of eSports is what Slovic (2010) describes as the ‘feeling of risk’.

These digital environments are able to manipulate and measure sensory input and alter the

perceived and actual levels of risk faced by players. Just like experimentalists do in a laboratory

setting, programmers would be able to test the perceptions of risk versus actual risk by control

or set the probability of certain events occurring and explore player reactions and decision-

making.

Conclusions

Throughout this book chapter we have tried to show how powerful sports data can be in

understanding questions that are at the core of behavioural economics. We sought to show how

47
scholars have advanced knowledge by finding new avenues through which they might improve

on previous papers that used sports data, by finding other areas of exploration within the sports

environment. We have also tried, without searching for completeness, to discuss throughout

the paper what else can be explored in the future in more detail. When exploring biases Camerer

and Weber (1999) raise an important point:

First, establishing systematic mistakes using naturally occurring data is very difficult. Of

course, this does not mean we should avoid such hard work and exploit the superior control of

the lab; it just means that the standard of proof for mistakes outside the lab is high, and should

be. It is also likely that important field anomalies will not be established by a single study, but

by a series of studies which build on earlier results. Behavioral economists have learned that

the best way to win an argument about the existence of systematic mistakes is to take

complicated rationalizations offered by critics seriously (no matter how cockamamie they are),

and collect more data to test them (p. 81).

Using sports data is a strategy to move towards that direction when working with naturally

occurring data. Sports data allow us to use a variety of different datasets and look at different

sport disciplines when exploring aspects in the area of behavioural economics. Our aim was to

show that sports data are a valuable tool of thought and exploration in the interplay between

speaking to theorists and searching for facts, which is particularly important when challenging

the status quo in a particular scientific field.

References

Aagaard, J. (2004). Inside the Chess Mind: How Players of All Levels Think about the Game. Gloucester
Publishers plc.
Allen, R. D., Hitt, A. M. and Greer, C. R. (1982) Occupational stress and perceived organizational
effectiveness in formal groups: An examination of stress level and stress type, Personal
Psychology, 35, 359–70.
Allison, P.D. (1999). Multiple Regression: A primer. Pine Forge Press.

48
Alter, A. L., and Oppenheimer, D. M. (2006). From a fixation on sports to an exploration of mechanism:
The past, present, and future of hot hand research. Thinking & Reasoning, 12(4), 431-444.
Arkes, H. R., and Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and
Human Decision Processes, 35(1), 124-140.
Avni, A. (2004). The Grandmaster’s Mind: A Look Inside the Chess Thinking-Process. Gambit
Publications Ltd.
Bar-Eli, M., Azar, O. H., Ritov, I., Keidar-Levin, Y., and Schein, G. (2007). Action bias among elite
soccer goalkeepers: The case of penalty kicks. Journal of Economic Psychology, 28(5), 606-621.
Baumeister, R. F. (1984) Choking under pressure: Self-consciousness and paradoxical effects of
incentives on skill performance, Journal of Personality and Social Psychology, 46, 610–20.
Baumeister, R. F. and Showers, C. J. (1986) A review of paradoxical performance effects: Choking
under pressure in sports and mental tests, European Journal of Social Psychology, 16, 361–83.
Beilock, S. L. and Carr, T. H. (2001) On the fragility of skilled performance: What governs choking
under pressure? Journal of Experimental Psychology: General, 130, 701–25.
Bem, D. J., Wallach, M. A., and Kogan, N. (1965). Group decision making under risk of aversive
consequences. Journal of Personality and Social Psychology, 1(5), 453–460.
Bénabou, R. (2013). Groupthink: Collective Delusions in Organizations and Markets, The Review of
Economic Studies, 80(2), 429–462.
Billings D., Papp D., Schaeffer J., Szafron D. (1998). Poker as a testbed for AI research. In: Mercer
R.E., Neufeld E. (Eds.), Advances in Artificial Intelligence. Canadian AI 1998. Lecture Notes in
Computer Science (Lecture Notes in Artificial Intelligence), vol 1418. Springer, Berlin,
Heidelberg.
Borland, J., Lee, L., and Macdonald, R. D. (2011). Escalation effects and the player draft in the AFL.
Labour Economics, 18(3), 371-380.
Bourne, L. E., & Yaroush, R. A. (2003). Stress and Cognition: A Cognitive Psychological Perspective,
National Aeronautics and Space Administration (N.A.S.A.).
Brouwer, T., and Potters, J. (2019). Friends for (almost) a day: Studying breakaways in cycling races.
Journal of Economic Psychology, 75, 102092.
Brown, N. and Sandholm, T. (2019). Superhuman AI for multiplayer poker. Science, 365(6456), 885-
890.
Bryson, A., Dolton, P., Reade, J. J., Schreyer, D., and Singleton, C. (2021). Causal effects of an absent
crowd on performances and refereeing decisions during Covid-19. Economics Letters, 198,
109664.
Burns, B. D. (2004). Heuristics as beliefs and as behaviors: The adaptiveness of the “hot hand”.
Cognitive Psychology, 48(3), 295-331.
Butler, J. L., and Baumeister, R. F. (1998). The trouble with friendly faces: Skilled performance with a
supportive audience. Journal of Personality and Social Psychology, 75(5), 1213-1230.

49
Camerer, C. F., and Weber, R. A. (1999). The econometrics and behavioral economics of escalation of
commitment: A re-examination of Staw and Hoang’s NBA data. Journal of Economic Behavior &
Organization, 39(1), 59-82.
Cao, Z., Price, J., and Stone, D. F. (2011). Performance under pressure in the NBA. Journal of Sports
Economics, 12(3), 231-252.
Caruso, R., Di Domizio, M., and Savage, D. A. (2017). Differences in National Identity, Violence and
Conflict in International Sport Tournaments: Hic Sunt Leones! Kyklos, 70(4), 511-545.
Chan, H. F., Savage, D. A., and Torgler, B. (2019). There and back again: Adaptation after repeated
rule changes of the game. Journal of Economic Psychology, 75, 102129.
Chase, W. G., and Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55-81.
Christakis, N. A. (2019). Blueprint: Evolutionary Origins of A Good Society. Little, Brown Spark.
Coates, J. (2012). The Hour Between Dog and Wolf: Risk-taking, Gut Feelings and the Biology of Boom
and Bust. Penguin Books.
Coates, D., and Oguntimein, B. (2010). The length and success of NBA careers: Does college
production predict professional outcomes. International Journal of Sport Finance, 5(1), 4-26.
Cohen-Zada, D., Krumer, A. and Shtudiner, Z. (2017). Psychological momentum and gender, Journal
of Economic Behavior & Organization, 135, 66-81. https://doi.org/10.1016/j.jebo.2017.01.009
Cohen-Zada, D., Krumer, A., Rosenboim, M., and Shapir, O. M. (2017). Choking under pressure and
gender: Evidence from professional tennis. Journal of Economic Psychology, 61, 176-190.
Coleman, J. S. (1998). Social Capital in the Creation of Human Capital, American Journal of Sociology,
94, S95-S120.
Cooper, J. (1989). The military and higher education in the USSR. The ANNALS of the American
Academy of Political and Social Science. 502: 108–119.
Darcy, S., Maxwell, H., Edwards, M., Onyx, J. and Sherker, S. (2014). More than a sport and volunteer
organisation: Investigating social capital development in a sporting organisation, Sport
Management Review, 17(4), 395-406.
Dasgupta, P. (1999), Economic progress and the idea of social capital, in Dasgupta, P. and Serageldin,
I. (Eds), Social Capital: A Multifaceted Perspective, World Bank.
de Groot, A. (1965). Thought and Choice in Chess. Mouton Publishers.
Diamond, J. (2019) The ‘Moneyball’ A’s Find a New Inefficiency: Other Teams’ Players, in the Wall
Street Journal, August 20, 2019. Available at https://www.wsj.com/articles/the-moneyball-as-
find-a-new-inefficiency-other-teams-players-11566322019.
Dobelli, R. (2013). The art of thinking clearly: better thinking, better decisions. Hachette UK.
Dohmen, T. J. (2008a). Do professionals choke under pressure?. Journal of Economic Behavior &
Organization, 65(3-4), 636-653.
Dohmen, T. J. (2008b). The influence of social forces: Evidence from the behavior of football referees.
Economic inquiry, 46(3), 411-424.

50
Dohmen, T., and Sauermann, J. (2016). Referee bias. Journal of Economic Surveys, 30(4), 679-695.
Duggan, M. and Levitt, S. D. (2002). Winning isn’t everything: Corruption in sumo wrestling. American
Economic Review, 92(5):1594–1605.
Dulleck, U., Ristl, A., Schaffner, M., and Torgler, B. (2011). Heart rate variability, the autonomic
nervous system, and neuroeconomic experiments. Journal of Neuroscience, Psychology, and
Economics, 4(2), 117-124.
Dulleck, U., Schaffner, M., and Torgler, B. (2014). Heartbeat and economic decisions: observing
mental stress among proposers and responders in the ultimatum bargaining game. PLoS One, 9(9),
e108218.
Dulleck, U., Fooken, J., Newton, C., Ristl, A., Schaffner, M., and Torgler, B. (2016). Tax compliance
and psychic costs: Behavioral experimental evidence using a physiological marker. Journal of
Public Economics, 134, 9-18.
Elias, N., and Dunning, E. (1966). Dynamics of sports groups with special reference to football. British
Journal of Sociology, 17(4), 388-402.
Elster, J. (1998). Emotions and economic theory. Journal of Economic Literature, 36(1), 47-74.
Emerson, J. W., Seltzer, M., and Lin, D. (2009). Assessing judging bias: An example from the 2000
Olympic Games. The American Statistician, 63(2), 124-131.
Eremin, O., Walker, M. B., Simpson, E., Heys, S. D., Ah-See, A. K., Hutcheon, A. W., Ogston, K. N.,
Sarkar, T. K., Segar, A. and Walker, L. G. (2009). Immuno-modulatory effects of relaxation
training and guided imagery in women with locally advanced breast cancer undergoing
multimodality therapy: A randomised controlled trial. Breast, 18(1), 17-25.
Flessas, K., Mylonas, D., Panagiotaropoulou, G., Tsopani, D., Korda, A., Siettos, C., Di Cagno, A.,
Evdokimidis, I. and Smyrnis, N. (2015). Judging the judges’ performance in rhythmic gymnastics.
Medicine & Science in Sports & Exercise, 47(3), 640-648.
Frank, R. H. and Cook, P. J. (1995). The Winner-Take-All Society. The Free Press.
Frank, R. H. (1984). Interdependent preferences and the competitive wage structure, RAND Journal of
Economics, 15(4), 510-520.
Frey, B. S., Schaffner, M., Schmidt, S. L., and Torgler, B. (2013). Do employees care about their
relative income position? Behavioral evidence focusing on performance in professional team sport.
Social Science Quarterly, 94(4), 912-932.
Fukuyama, F. (2003). Social capital and civil society.In Ostrom, E. and Ahn, T.K. (Eds.), Foundations
of Social Capital. Edward Elgar.
Garicano, L., Palacios-Huerta, I., and Prendergast, C. (2005). Favoritism under social pressure. Review
of Economics and Statistics, 87(2), 208-216.
Gauriot, R. and Page, L. (2018). Psychological momentum in contests: The case of scoring before half-
time in football, Journal of Economic Behavior & Organization, 149, 137-168.

51
Gigerenzer, G. (1996). On narrow norms and vague heuristics: A reply to Kahneman and Tversky.
Psychological Review, 100, 592-596.
Gigerenzer, G. (2000). Adaptive thinking: Rationality in the real world. Oxford University Press.
Gigerenzer, G. (2004). Striking a blow for sanity in theories of rationality. In: M. Augier and J. G.
March (Eds.), Models of a man: Essays in memory of Herbert A. Simon. MIT Press, 389-409.
Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of
random sequences. Cognitive Psychology, 17(3), 295-314.
Goff, B. L., and Tollison, R. D. (Eds.). (1990). Sportometrics. Texas A&M University Press.
Groothuis, P. A., and Hill, J. R. (2004). Exit discrimination in the NBA: A duration analysis of career
length. Economic Inquiry, 42(2), 341-349.
Groysberg, B., Nanda, A., and Nohria, N. (2004). The risky business of hiring stars. Harvard Business
Review, 82(5), 92-101.
Hackinger, J. (2019). Ignoring millions of euros: Transfer fees and sunk costs in professional football.
Journal of Economic Psychology, 75, 102114.
Hamilton, D. R. (2018). How Your Mind Can Heal Your Body, Hay House.
Harb-Wu, K., and Krumer, A. (2019). Choking under pressure in front of a supportive audience:
Evidence from professional biathlon. Journal of Economic Behavior & Organization, 166, 246-
262.
Heiniger, S., and Mercier, H. (2019). Judging the judges: A general framework for evaluating the
performance of international sports judges. arXiv preprint arXiv:1807.10055.
Hickman, D. C., and Metz, N. E. (2015). The impact of pressure on performance: Evidence from the
PGA TOUR. Journal of Economic Behavior & Organization, 116, 319-330.
Hill, A. (2018). Why groupthink never went away. Financial Times, May 7th, 2018. Available from
https://www.ft.com/content/297ffe7c-4ee4-11e8-9471-a083af05aea7
Hill, R. A., and Barton, R. A. (2005). Red Enhances Human Performance in Contents. Nature 435, 293–
293.
Howell, R. (1975). The USSR: Sport and politics intertwined. Comparative Education, 11(2): 137–145.
Humphreys, B. and Ruseski, J. (2011). Socio-Economic Determinants of Adolescent Use of
Performance Enhancing Drugs: Evidence from the YRBSS, The Journal of Socioeconomics, 40(2),
208–216.
Ilie, A., Ioan, S., Zagrean, L. and Moldovan. M. (2008). Better to be Red Than Blue in Virtual
Competition. Cyber Psychology and Behaviour, 11(3), 375–377.
Iso-Ahola, S. E., and Dotson, C. O. (2016). Psychological Momentum-A Key to Continued Success.
Frontiers in Psychology, 7, 1328.
Jamal, M. (1984). Job stress and job performance controversy: An empirical assessment,
Organizational Behavior and Human Performance, 33, 1–21.
Janis, I. L. and Mann, L. (1977) Decision Making, The Free Press.

52
Janis, I. L. (1972). Victims of groupthink: A psychological study of foreign-policy decisions and fiascos.
Houghton Mifflin.
Kahn, L. M. (2000). The sports business as a labor market laboratory. Journal of Economic
Perspectives, 14(3), 75-94.
Kahneman, D., and Tversky, A., (1979). Prospect theory: An analysis of decision under risk.
Econometrica 47(2), 263–291.
Kawachi I., Subramanian S., and Kim D. (2008). Social Capital and Health. In: Kawachi I.,
Subramanian S., Kim D. (Eds.) Social Capital and Health. Springer. New York.
Kausel, E. E., Ventura, S., and Rodríguez, A. (2019). Outcome bias in subjective ratings of
performance: Evidence from the (football) field. Journal of Economic Psychology, 75, 102132.
Kay, T. and Bradbury, S. (2009) Youth sport volunteering: developing social capital?, Sport, Education
and Society, 14(1), 121-140.
Keefer, Q. A. (2015a). Performance feedback does not eliminate the sunk-cost fallacy: Evidence from
professional football. Journal of Labor Research, 36(4), 409-426.
Keefer, Q. A. (2017). The sunk-cost fallacy in the National Football League: Salary cap value and
playing time. Journal of Sports Economics, 18(3), 282-297.
Keltner, Da. and Lerner, J. S. (2010). Emotion. In: D. Gilbert, S. Fiske, and G. Lindsey (Eds.). The
Handbook of Social Science, Vol. 1. Wiley, pp. 317-352.
Keinan, G. (1987) Decision making under stress: Scanning of alternatives under controllable and
uncontrollable threats, Journal of Personality and Social Psychology, 52, 639–44.
Kho, A. Y., Liu, K. P., Chung, R. C. (2014). Meta-analysis on the effect of mental imagery on motor
recovery of the hemiplegic upper extremity function. Australian Occupational Therapy Journal,
61(2), 38-48.
Kim, A. C. H., Ryu, J., Lee, C. et al. (2021) Sport Participation and Happiness Among Older Adults: A
Mediating Role of Social Capital. Journal of Happiness Studies, 22, 1623-1641
Koehler, J. J., & Conley, C. A. (2003). The “hot hand” myth in professional basketball. Journal of Sport
and Exercise Psychology, 25(2), 253-259.
Kraus, S., Meier, F., Niemand, T., Bouncken, R. B. and Ritala, P. (2018). In search for the ideal
coopetition partner: an experimental study. Review of Managerial Science, 12, 1025–1053.
Krumer, A. (2020). Pressure versus ability: Evidence from penalty shoot-outs between teams from
different divisions. Journal of Behavioral and Experimental Economics, 89, 101578.
Kumar, H., Manoli, A. E., Hodgkinson, I. R. and Downward, P. (2018). Sport participation: From
policy, through facilities, to users’ health, well-being, and social capital, Sport Management
Review, 21(5), 549-562.
Leeds, D. M., Leeds, M. A., and Motomura, A. (2015). Are sunk costs irrelevant? Evidence from
playing time in the National Basketball Association. Economic Inquiry, 53(2), 1305-1316.

53
Lehner, P., Seyed-Solorforough, M., O’Connor, M. F., Sak, S. and Mullin, T. (1997) Cognitive biases
and time stress in team decision making, Transactions on Systems, Man, and Cybernetics Part A:
Systems and Humans, 27, 698–703.
Levitt, S. D. (2002). Testing the economic model of crime: The national hockey league's two-referee
experiment. Contributions in Economic Analysis & Policy, 1(1).
Lewis, M. (2017). The Undoing Project: A Friendship that Changed Our Minds. W. W. Norton &
Company.
Lewis, M. (2004). Moneyball: The Art of Winning an Unfair Game. W. W. Norton & Company.
Locke, E. A. (1968). Toward a theory of task motivation and incentives. Organizational Behavior and
Human Performance, 3(2), 157–189.
Locke, E. A., Latham, G. P., Smith, K. J. and Wood, R. E. (1990). A Theory of Goal Setting & Task
Performance. Prentice Hall.
Loewenstein, G. (2005). Hot-cold empathy gaps and medical decision making. Health Psychology,
24(4), S49–S56.
Nowak, M., and Highfield, R. (2011). Supercooperators: Altruism, evolution, and why we need each
other to succeed. Simon and Schuster.
Macintyre, Alison, Ho Fai Chan, Markus Schaffner, and Benno Torgler (2021). National Pride and Tax
Compliance: A Laboratory Experiment Using a Physiological Marker. CREMA Working Paper
No. 2021-07. Center for Research in Economics, Management and the Arts (CREMA).
McCormick, R. E., and Tollison, R. D. (1984). Crime on the court. Journal of Political Economy, 92(2),
223-235.
Meglino, B. M. (1977) Stress and performance – are they always incompatible, Supervisory
Management, 22, 2–13.
Meichenbaum, D. (2007), Stress inoculation training: a preventative and treatment approach. In: Lehrer,
M., Woodfolk, R.L. and Slime, W.S. (Eds), Principles and Practices of Stress Management.
Guilford Press.
Merkel, S., Chan, H. F., Schmidt, S. L., and Torgler, B. (2021). Optimism and positivity biases in
performance appraisal ratings: Empirical evidence from professional soccer, forthcoming in:
Applied Psychology.
Myers, T. D., Balmer, N. J., Nevill, A. M., and Al Nakeeb, Y. (2006). Evidence of nationalistic bias in
muaythai. Journal of Sports Science & Medicine, 5(CSSI), 21-27.
Nalebuff, B. J., and Brandenburger, A. M. (1997). Co-opetition: Competitive and Cooperative Business
Strategies for the Digital Economy. Strategy & Leadership, 25(6), 28–33.
Newborn, M. (1975). Computer Chess. ACM Monograph Series. Academic Press.
Newell, Al. (1954). The Chess Machine: An Example of Dealing With a Complex Task by Adaptation,
P-620, Rand Corporation.
Newell, A., and Simon, H. A. (1972). Human problem solving. Prentice-Hall, Inc.

54
Paldam, M. (2000). Social capital: One or many? Definition and measurement, Journal of Economic
Surveys, 14(5), 629-653.
Page, L., & Coates, J. (2017). Winner and loser effects in human competitions. Evidence from equally
matched tennis players. Evolution and Human Behavior, 38(4), 530-535.
Pedace, R., and Smith, J. K. (2013). Loss aversion and managerial decisions: Evidence from major
league baseball. Economic Inquiry, 51(2), 1475-1488.
Pettersson-Lidbom, P., and Priks, M. (2010). Behavior under social pressure: Empty Italian stadiums
and referee bias. Economics Letters, 108(2), 212-214.
Piatti, M., Savage, D. A. and Torgler, B. (2012), The Red Mist? Red shirts, success and team sports,
Sport in Society, 15(9), 1209-1227.
Picknell, D. (2019). What Is Esports and How Did it Become a $1 Billion Industry? Learning Hub
Article August 20th, 2019. Accessed from https://learn.g2.com/esports
Pollak, R. A. (1976). Interdependent Preferences. The American Economic Review, 66(3), 309-320.
Portes, A. (1998). Social capital: its origins and applications in contemporary sociology, Annual Review
of Sociology, 24, 1-24.
Postlewaite, A. (1998). Social Status, Norms and Economic Performances: The social basis of
interdependent preferences, European Economic Review, 42, 779–800.
Praetorius, A. S. and Görlich, D. (2020) How Avatars Influence User Behavior: A Review on the
Proteus Effect in Virtual Environments and Video Games, FDG ‘20: International Conference on
the Foundations of Digital Games 49, 1–9.
Rasskin-Gutman, D. (2009). Chess metaphors: Artificial intelligence and the human mind. MIT Press.
Ratan, R., Beyea, D., Li, B. J. and Graciano, L. (2020). Avatar characteristics induce users’ behavioral
conformity with small-to-medium effect sizes: A meta-analysis of the proteus effect, Media
Psychology, 23(5), 651-675.
Riordan, J. (1993). The rise and fall of soviet Olympic champions. Olympika: The International Journal
of Olympic Studies. 2: 25–44.
Savage, D. A. and Torgler, B. (2012). Nerves of steel? Stress, work performance and elite athletes.
Applied Economics. 44(19): 2423–2435.
Schaffner, M., and Torgler, B. (2008). Meet the Joneses: An Empirical Investigation of Reference
Groups in Relative Income Position Comparisons. CREMA Working Paper No. 2008-13. Center
for Research in Economics, Management and the Arts (CREMA).
Schmidt, S. L. (Ed.). 21st Century Sports: How Technologies Will Change Sports in the Digital Age.
Springer.
Schultz, D. P. (1966) An experimental approach to panic behavior, Group Psychology Branch.
Schwert, G. W. (2003). Anomalies and Market Efficiency (Chapter 15). In: G. M. Constantinides, M.
Harris, and R. M. Stulz (Eds.), Handbook of Economics of Finance. Elsevier, pp. 939-974.
Simon, H. (1983). Reason in human affairs. Stanford University Press.

55
Simon, H. A. (1996). Models of my life. MIT press.
Skinner, J., Zakus, D. H. and Cowell, J. (2008). Development through Sport: Building Social Capital
in Disadvantaged Communities, Sport Management Review, 11(3), 253-275.
Slovic, P. (2010). The feeling of risk: New perspectives on risk perception. Routledge.
Staw, B. M., and Hoang, H. (1995). Sunk costs in the NBA: Why draft order affects playing time and
survival in professional basketball. Administrative Science Quarterly, 40, 474-494.
Stoner, J. A. (1959). A comparison of individual and group decisions involving risk, Master’s Thesis,
Antioch College, https://dspace.mit.edu/bitstream/handle/1721.1/11330/33120544-MIT.pdf
Sullivan, S. E. and Bhagat, R. S. (1992) Organizational stress, job satisfaction and job performance:
where do we go from here? Journal of Management, 18, 353–74.
Thaler, R. H. (2015). Misbehaving: The Making of Behavioral Economics. W. W. Norton and
Company.
Thaler, R. H., Johnson, and E. J., (1990). Gambling with the house money and trying to break even:
The effects of prior outcomes on risky choice. Management Science, 36(6), 643–660.
Toma, M. (2017). Missed shots at the free-throw line: Analyzing the determinants of choking under
pressure. Journal of Sports Economics, 18(6), 539-559.
Torgler, B. (2004). The economics of the FIFA Football Worldcup. Kyklos, 57(2), 287-300.
Torgler, B. (2009). Economics of sports: A note to this special issue. Economic Analysis and Policy,
39(3), 333.
Torgler, B. (2016). Can Tax Compliance Research Profit from Biology?, Review of Behavioral
Economics, 3, 113-144, 2016.
Torgler, B. (2019). Opportunities and challenges of portable biological, social, and behavioral sensing
systems for the social sciences. In G. Foster (Ed.), Biophysical measurement in experimental social
science research. Academic Press, pp. 197-224.
Torgler, B. (2020). Big Data, Artificial Intelligence, and Quantum Computing in Sports. In: S. L.
Schmidt (Ed.), 21st Century Sports: How Technologies Will Change Sports in the Digital Age.
Springer, pp. 153-173,
Torgler, B. (2021a). Symbiotics> Economics? CREMA Working Paper No. 2021-15. Center for
Research in Economics, Management and the Arts (CREMA).
Torgler, B. (2021b). The Power of Public Choice in Law and Economics. CREMA Working Paper No.
2021-04. Center for Research in Economics, Management and the Arts (CREMA).
Van Ours, J. C., and Van Tuijl, M. A. (2016). In-season head-coach dismissals and the performance of
professional football teams. Economic Inquiry, 54(1), 591-604.
Wiederhold, B.K. (2013). Avatars: Changing Behavior for Better or for Worse? Cyberpsychology,
Behavior, and Social Networking, 16(5), 319-320.
Walseth, K. (2008). Bridging and bonding social capital in sport – experiences of young women with
an immigrant background, Sport, Education and Society, 13(1), 1-17.

56
Wilson, D. S. (2019). This view of life: Completing the Darwinian revolution. Pantheon Books.
Witt, R. (2005). Do players react to sanction changes? Evidence from the English Premier League.
Scottish Journal of Political Economy, 52(4), 623-640.
Woolcock, M. and Narayan, D. (2000), Social capital: Implications for development theory, research
and policy, The World Bank Research Observer, 2, 225-249.
Wright, P. (1974) The harassed decision maker: Time pressures, distractions and the use of evidence,
Journal of Applied Psychology, 59, 555–61.
Yerkes, R. and Dodson, J. D. (1908) The relationship of stimuli to rapidity of habit formation, Journal
of Comparative Neurological Psychology, 18, 459–82.
Zeelenberg, M., Van den Bos, K., Van Dijk, E., and Pieters, R. (2002). The inaction effect in the
psychology of regret. Journal of Personality and Social Psychology, 82(3), 314-327.
Zitzewitz, E. (2006). Nationalism in winter sports judging and its lessons for organizational decision
making. Journal of Economics & Management Strategy, 15(1), 67-99.
Zitzewitz, E. (2014). Does transparency reduce favoritism and corruption? Evidence from the
reform of figure skating judging. Journal of Sports Economics, 15(1), 3-30.

57