O 91^1
¿2
CAN GAMBLERS BEAT RANDOMNESS ? AN EXPERIMENTAL STUDY ON
SPORT BETTING
Mémoire présenté
à la Faculté des études supérieures de l’Université Laval
pour l’obtention du grade de maître es psychologie (M.Ps.)
École de psychologie
FACULTÉ DES SCIENCES SOCIALES UNIVERSITÉ LAVAL
DÉCEMBRE 2002
Bien que les habiletés ne soient pas considérées pertinentes dans le cadre des jeux de hasard, peu d’études ont évalué si les habiletés des parieurs sportifs chevronnés sont une manifestation de l’illusion de contrôle. Cette recherche examine donc : (1) si les experts en pari sportif sur le hockey peuvent obtenir des pronostics plus précis que le hasard, (2) s’ils peuvent obtenir des gains monétaires plus élevés que le hasard et (3) sur quelles informations les parieurs se basent lorsqu’ils misent. Pour ce faire, 30 participants ont communiqué à 6 reprises leurs paris sportifs effectués auprès de la loterie d’État et ont rempli un questionnaire sur le pari sportif. Malgré un taux de précision plus élevé que le hasard, les gains des experts ne sont pas significativement plus élevés que ceux du hasard. Il est suggéré que les informations utilisées par les parieurs, ainsi que les quasi-succès (niveau de précision) renforcent la perception d’expertise et l’illusion de contrôle des parieurs.
Michael Cantinotti Robert Ladouceur, Ph.D
Although skills are not considered relevant in chance governed activities, only few studies have assessed to which extent sport expert skills in wagering are a manifestation of the illusion of control. Thus, the present paper examines: (1) if expert hockey bettors can make better predictions than random selection, (2) if expert hockey bettors can achieve greater monetary gains than what can be expected from chance, and (3) what kind of information and strategies hockey gamblers rely on when betting. Accordingly, 30 participants were asked to report their state lottery hockey bets on 6 occasions. They also filled in a questionnaire on sports wagering. Despite a rate accuracy greater than chance, the monetary gains of expert hockey gamblers are not significantly higher than what can be expected by chance. It is suggested that the information used by bettors, along with near-misses (level of precision), reinforce their perception of expertise and their illusion of control.
Robert Ladouceur, Ph.D Directeur de recherche Michael Cantinotti
Pour avoir accepté de miser sur un étudiant de la lointaine Helvétie, pour la confiance offerte tout au long des projets auxquels il m’a proposé de participer, mais aussi pour ses critiques constructives et son soutien, je désire remercier mon directeur de recherche, Robert Ladouceur.
Ma reconnaissance va également à Christian Jacques et à Serge Sévigny, qui m’ont offert leur temps, leur humour et leurs réflexions pertinentes. Je tiens également à remercier Jean-Bernard Dupuy, pour sa constante disponibilité, sa bonne humeur et son appui en anglais pour la traduction de l’article. Quant au choc culturel, c’est à toute l’équipe du lab que je le dois... J’espère pouvoir perfectionner mon jouai auprès de vous encore un certain temps !
Mon Père, Dino, et ma Mère, Christine, m’ont appris la valeur du travail, des petits bonheurs de la vie et m’ont insufflé le goût de la découverte. Vingt-cinq ans plus tard, j’espère que je saurais vous remercier un jour pour tout ce que vous m’avez donné.
Un grand merci à Olivia également, pour son temps et son soutien. Sans ton aide, ces milliers de données n’auraient pu être compilées !
Enfin, ma reconnaissance va à l’État du Valais, la Fondation Max-Husmann (Zürich), ainsi qu’à la Fondation Suzanne et Hans Biäsch (Zürich), qui m’ont offert leur soutien financier. Leur aide a rendu ce rêve québécois possible.
Page IV V vii 1 Résumé Abstract Avant-propos Table des matières Liste des tableaux Section I :
INTRODUCTION GÉNÉRALE Section II :
CAN GAMBLERS BEAT RANDOMNESS ?
AN EXPERIMENTAL STUDY ON SPORT BETTING (Article) 8
1. ABSTRACT 10 2. INTRODUCTION 11 3. METHOD 16 3.1. Participants 16 3.2. Material 17 3.3. Procedure 17 4. RESULTS 19 4.1. Wagers’ accuracy 19 4.2. Monetary results 20
4.3. Experts’ strategies and ways of playing 21
22 5. DISCUSSION
30 Page 6. REFERENCES 34 7. AUTHOR NOTE 8. FOOTNOTES 35 9. TABLES 36 Section III : CONCLUSION GÉNÉRALE 39
Liste des ouvrages cités 41
Annexes
Annexe A : Formulaire de consentement 48
Annexe B : Questionnaire sur Mise-O-Jeu Hockey 51
Annexe C : Questionnaire SOGS (version téléphonique francophone) 56 Annexe D : Échantillon de programme et de fiche Mise-O-Jeu 60
35
36 Page Average accuracy of bets
Average amount of money ($) returned on bets wagered by experts and by random selection Sources of information utilized by participants Table 1 :
Table 2 :
37 Table 3 :
Historiquement, mis à part le pari mutuel sur les chevaux, l’émergence en Amérique du nord des jeux de pari sportif légaux est relativement récente. Excepté au Nevada, les premières tentatives d’instauration d’un système de loterie sur les événements sportifs se soldèrent par un échec. Smith (1992) évoque deux
principales raisons pour expliquer la réticence du public envers ces jeux : D’une part, les ligues de sport professionnel s’opposèrent à la tenue de ces paris, arguant que cela pourrait nuire à leur réputation; d’autre part, le fait de miser de l’argent sur une activité considérée noble et porteuse de valeurs comme le sport semblait immoral pour une partie du public, étant donné les risques de tricherie.
Aux États-Unis, le gouvernement fédéral a récemment renforcé son arsenal judiciaire (U.S. Senate Committee on Commerce, Science, and Transportation, 2001), afin de limiter au maximum la prolifération du pari sportif, sur la base des conseils prodigués par la National Gambling Impact Study Commission (1999). Cette dernière instance soutient que le pari sportif légalisé véhicule le message que le sport est plus une question d’argent qu’une question d’achèvement personnel, que le pari sportif menace l’intégrité du sport et qu’il mine la confiance du public.
Conséquemment, seuls deux États proposent actuellement du pari sportif légal aux Etats-Unis (Oregon et Nevada), bénéficiant de lois étatiques qui leur sont propres et antérieures aux lois édictées par le fédéral (National Gambling Impact Study
Commission, 1999).
Au Canada, contrairement aux pays européens comme l’Angleterre
(Department for Culture, Media and Sport, 2001), la France (Sénat français, 2002) ou l’Italie (Eurispes, 2000), les paris sportifs légaux ne se sont développés que
plusieurs sports (base-ball, football, hockey, soccer), tant au niveau nord-américain qu’en dehors du continent. La fréquence des parties à disposition a constamment augmenté depuis !’introduction en 1990 de Mise-O-Jeu, l’un des paris sportifs légaux québécois (Loto-Québec, 2001). Le fait que ce nouveau jeu ait franchi le seuil du demi-milliard de dollars de ventes cumulées (Loto-Québec, 2001) indique que cette nouvelle forme de jeu d’argent légal a atteint un groupe cible de joueurs.
Parmi les jeux de hasard et d’argent, le pari sportif semble occuper une place à part. Alors que les chercheurs s’accordent sur le fait que les jeux basés sur le hasard sont imprédictibles étant donné leur nature aléatoire, il n’en va pas de même pour le pari sur les événements sportifs. Certains chercheurs (Allcock, 1987; Burger, 1991; Rogers, 1998) suggèrent que dans ce cas, des éléments de connaissance ou des habiletés particulières sont nécessaires pour réaliser des paris efficaces.
À l’inverse, Wood (1992), en se basant sur la saison 1988 de base-ball, suggère que les parieurs ont tendance à surestimer leur capacité à prédire les résultats des parties : seul 1.7% de la variance des résultats des parties était
imputable aux trois meilleurs prédicteurs (lancer gagnant, avantage du terrain, fiche des victoires et des défaites de chaque équipe). Le fait de prétendre pouvoir prédire les résultats de base-ball en se basant sur ces variables peut donc être assimilé à une illusion cognitive. De façon analogue, Ladouceur, Giroux et Jacques (1998) ont démontré qu’au pari hippique, les habiletés des joueurs ne sont qu’une manifestation de l’illusion de contrôle. Dans cette étude, il est apparu que bien que les parieurs obtiennent un niveau de précision plus élevé que le hasard, ils ne parviennent pas à obtenir des gains supérieurs au hasard.
Ces résultats ne peuvent cependant être généralisés d’office aux autres paris sur événements sportifs. En effet, la structure (ou les règles) change selon les jeux. Il apparaît donc nécessaire de répliquer l’étude de Ladouceur et al. (1998) sur d’autres sports que les courses de chevaux, afin de vérifier quel est le rôle réel des habiletés aux paris sportifs. De plus, il serait nécessaire d’évaluer quels sont les facteurs que les parieurs prennent en considérations pour estimer qu’ils ont un certain contrôle (ou certaines connaissances) quant à l’issue des événements sportifs.
Langer (1975) considère qu’il y a illusion de contrôle lorsqu’une personne estime que sa probabilité subjective de succès à un événement donné est supérieure à la probabilité objective. Bien que les joueurs parviennent à obtenir le cheval
gagnant plus souvent que le hasard, leur résultat monétaire est équivalent au tirage aléatoire. Or, le pari sur événements sportifs est avant tout un jeu d’argent et son attrait réside essentiellement dans la possibilité d’obtenir des gains monétaires, bien que d’autres éléments motivationnels soient également présents (Keren et
Wagenaar, 1988). Pour cette raison, il est justifié de considérer que dans une
perspective d’évaluation de l’efficacité des paris, ce n’est pas tant le taux de précision des paris qui importe, mais bien plus le taux de retour monétaire obtenu par les
joueurs.
À Mise-O-Jeu, les participants doivent prédire l’issue d’événements sportifs. Sur la base d’un programme publié par l’organisateur, de 3 à 6 parties doivent être sélectionnées et le résultat (victoire de l’équipe visiteuse, partie nulle ou victoire de l’équipe receveuse) indiqué sur une fiche de sélection. Pour chacune des sélections, le participant doit indiquer le montant misé. Le résultat monétaire potentiel est calculé en multipliant le montant misé avec les cotes des résultats associées aux prédictions
du parieur. Ces cotes sont inscrites sur le programme des parties. Pour obtenir un gain monétaire, la sélection doit être entièrement correcte. À titre d’exemple, un joueur qui obtient 5 résultats de parties corrects sur 6 ne gagne rien.
Mise-O-Jeu est-il un jeu d’habileté ou un jeu de hasard? Ou encore, est-ce un
jeu dans lequel habileté et hasard s’entremêlent? Dans ce dernier cas de figure, quelle part peut-on attribuer aux habiletés et au hasard? La réponse à ces questions est loin d’être évidente. Keren et Wagenaar (1988) avancent qu’il est impossible d’utiliser des critères mesurables et objectifs pour distinguer les jeux d’adresse de ceux de hasard de façon définitive. Selon le raisonnement de ces auteurs, il est également impossible de quantifier objectivement la part de hasard et d’habileté qui prévaut dans les jeux mixtes. Il semble toutefois que cette quantification ne soit pas vraiment essentielle : ce sont plutôt (a) la structure financière du jeu (l’espérance de gain est elle négative?), (b) la présence d’une part de hasard et (c) la prise de risque de la part des joueurs qui importent dans le cadre d’un jeu d’argent (Keren &
Wagenaar, 1988). Un jeu d’habileté sur lequel des mises sont effectuées et où l’espérance mathématique du taux de retour est négative est donc avant tout un jeu de hasard et d’argent.
En outre, la controverse ne s’est pas limitée au niveau scientifique. Au niveau judiciaire, la partie VII (Maisons de désordre, jeux et paris) du Code criminel
(Ministère de la Justice du Canada, 2001) couvre explicitement les jeux de hasard. Au cours des années 1980, la lutte pour le marché des jeux de hasard qui eut lieu entre les provinces et le gouvernement fédéral du Canada amena ce dernier à
intenter un procès à Loto-Québec pour la création d’un pari sportif sur le hockey dont le but était de prédire le résultat final de treize parties. Le gouvernement fédéral
reprochait à la société de loterie de faire usage d’un jeu d’argent basé sur les
habiletés, donc illégal. Dans un jugement qui allait conditionner !’interprétation future du Code criminel, la Cour supérieure du Québec indique que ce type de jeu ne contrevient pas à la loi, étant donné qu’il « est un jeu mixte de hasard et d’habileté » (Société Québécoise d’information Juridique, 1983).
S’il est effectivement indubitable qu’un joueur de pari sportif puisse avoir de bonnes connaissances du sport sur lequel il mise, il est par contre moins certain que ces éléments lui soient utiles pour obtenir un résultat monétaire excédentaire en comparaison des montants misés. Il est frappant de constater que malgré l’évidence que le pari sportif est un jeu où il y a un élément de hasard (sinon, il serait déclaré illégal) et pour lequel le taux de retour de la majorité des joueurs est négatif (quel intérêt y aurait-il pour une société de loterie de gérer un jeu générant des déficits?), le mythe du parieur habile et gagnant se perpétue. À ce sujet, la recherche menée par Gilovich (1983) comporte des éléments révélateurs : la persistance à miser de l’argent sur des événements imprédictibles est due à l’évaluation cognitive biaisée des résultats par les parieurs. En résumé, il apparaît que les joueurs tendent à considérer leurs résultats positifs comme évidents, les estimant révélateurs de leurs habiletés, alors que la portée des résultats négatifs est diminuée, les joueurs les attribuant à des facteurs externes incontrôlables (ex. : malchance, chance de l’équipe adverse).
L’objectif de la présente recherche consiste donc à vérifier si les habiletés et les connaissances dont les parieurs se servent sont réelles ou si elles sont plutôt une manifestation de l’illusion de contrôle que l’on retrouve dans la plupart des jeux de hasard et d’argent. Pour ce faire, nous avons décidé de nous baser sur les paris sur
le hockey. Ce sport est, pour des raisons historiques, culturelles et
environnementales, le plus suivi au Québec. La probabilité de pouvoir recruter des personnes se disant « expertes » dans le domaine du pari sur le hockey était donc vraisemblablement la plus favorable.
Running head: CAN GAMBLERS BEAT RANDOMNESS ?
Can Gamblers Beat Randomness ? An Experimental Study on Sport Betting Michael Cantinotti, Robert Ladouceur and Christian Jacques
Abstract
Although skills are not considered relevant in chance governed activities, only few studies have assessed to which extent sport expert skills in wagering are a
manifestation of the illusion of control. Thus, the present paper examines: (1) if expert hockey bettors can make better predictions than random selection, (2) if expert
hockey bettors can achieve greater monetary gains than what can be expected from chance, and (3) what kind of information and strategies hockey gamblers rely on when betting. Accordingly, 30 participants were asked to report their state lottery hockey bets on 6 occasions. They also filled in a questionnaire on sports wagering. Despite a rate accuracy greater than chance, the monetary gains of expert hockey gamblers are not significantly higher than what can be expected by chance. It is suggested that the information used by bettors, along with near-misses (level of precision), reinforce their perception of expertise and their illusion of control. KEYWORDS: sport, betting, gambling, strategy, illusion of control.
Can Gamblers Beat Randomness ? An Experimental Study on Sport Betting Until ten years ago, the effort of most North American Lotteries to introduce legal wagers on sports was unsuccessful. According to Smith (1992), this failure can be explained as follows: First of all, professional sport associations opposed
themselves to sports wagering, fearing that it would damage their reputation. This argument is linked to the fact that the public generally considers sports as a basis for learning about life, particularly for youths, and that sport should also be preserved from cheating caused by gambling. As a consequence, politicians are reluctant to consider the legalization of sports wagering. Moreover, for a $2 bet, the potential winnings are low when compared with other traditional lotteries, discouraging gamblers dreaming of fortune. The situation has gradually changed, especially in Canada, with the legalisation of sports betting from governments hoping to take in additional revenues and with the marketing campaigns of sport lotteries, interested too in targeting a new group of buyers (Smith, 1992).
In Canada, sports wagering is legal in all the provinces and is run by state lotteries. On the other hand, sports wagering is illegal in most states in the United States. Sport wagering is allowed in Nevada, Oregon, Delaware and Montana only (however, the two latter states don’t presently offer those kinds of games ), although many Americans seem to be unaware of that fact. What is more, illegal sports betting was evaluated between $80 billion and $380 billion in the United States, in 1999 (National Gambling Impact Study Commission, 1999). This suggests that sports wagering is “the most widespread and popular form of gambling in America” (National Gambling Impact Study Commission, 1999, p. 2.14).
When the first legal sport books were introduced by lotteries, the number of betting possibilities were usually restricted to one sport, like in Delaware in 1976 (see Smith, 1992). However, in the last ten years, there has been a progressive increase in the number of sports events on which gamblers can legally bet in Canada. For example, in Quebec, people can now gamble on popular sports like hockey, football, baseball and even soccer. They can choose games taking place in North America or abroad.
An important question remains about the nature of sports betting. Contrary to horse racing, where the skills and knowledge of the players are of no help for winning money (Ladouceur, Giroux & Jacques, 1998), the same has not been clearly
established in other sports like football, basketball, tennis and hockey. The main question under consideration in the present context is the following one: Are the skills and knowledge used by the gamblers real or illusory? In order to answer this
question, it is necessary to know if sports betting is based on the principle of chance and randomness, as in a traditional lottery game, or not.
Apart from the wagers on horses (Ladouceur et al., 1998), the relation between gamblers’ skills and monetary outcomes in sports betting has only been scarcely studied. It is well known since the Langer’s study (1975) that the various kinds of information gamblers rely on when they bet on events determined by randomness is only reinforcing their illusion of control. Langer (1975) defines the illusion of control as follows: “an expectancy of a personal success probability inappropriately higher than the objective probability would warrant” (p. 313). For example, when tossing a coin, someone who contends that the probability of a head is more than 50% if the
chance governs traditional lotteries, like coin tossing or dice playing (unless the dice are loaded), it could not be excluded that the information on which sport experts rely can to a certain extent be useful.
Indeed, some studies conclude that certain factors like home play and ranking significantly affect the game results (Vergin & Sosik, 1999; Boulier & Stekler, 1999). Moreover, some researchers suggest that skills could be useful for successefull betting on sports events (Allcock, 1987; Burger, 1991; Rogers, 1998). Professional sports handicappers take into account a lot of facts about teams and games (wounds of key players, last results of teams, home field/ice advantage, odds,1 etc.) to
determine the best wagering opportunities (York, 2002). Is this information really useful for sports gamblers or does it give rise to an illusion of control ?
Most people know that outcomes of roulette or lotteries are based on the notion of chance or randomness. Although no one can successfully predict the
outcome of the game, many gamblers try to predict the outcome of the game as if the results were linked together with previous results (Ladouceur, Dube, Giroux,
Legendre, & Gaudet, 1995). Often, they try to predict the outcome of games by using a variety of illusory strategies, for example by basing themselves on past results when making their next bet. People are reluctant to accept that the rule stipulating that we can learn from the past cannot be extended to gambling (Wood, 1992).
Contrary to lotteries, in sports events very few people indicate that the
outcome is only a matter of chance. For this reason, it is reasonable to suppose that sports bettors will have a deterministic view of sport events, instead of a probabilistic one. We have learned from experience that to achieve good results in a sport
fact to the conclusion that when we obtain a winning bet at sports wagering that is mainly because we made a well-informed wager. Is this statement supported at least to a certain extent or is it completely erroneous? The process of attributing causality between events when they are successive is a common form of reasoning (Dawes, 1988). To view the world we live in as a predictable one is quite understandable and adaptative (Wortman, 1976). Unfortunately, this perception is completely unfounded when the results follow a random sequence. Because a lot of information linked to games is available for sports bettors, it would be interesting to know if this information is of any help for determining the games’ outcomes. If this were not the case, that would mean that perceived skills in sport betting only are a manifestation of the illusion of control often found in most gambling activities.
The legal sport betting game in Quebec is called Mise-O-Jeu (literally “Bet-On- Game”).2 When betting at Mise-O-Jeu, participants must select from 3 to 6 games3 of the program (published in most newspapers) for each bet and make their predictions by filling out a selection slip. Programs are of variable length (number of games to bet on), depending on the number of matches which take place during the week. The different outcomes of the games that can be wagered are: (a) visiting team wins, (b) tie game, or (c) home team wins. Participants can calculate their potential winnings by multiplying the amount wagered by the product of the odds quoted for each of their predictions. Winnings occur only if all the predictions are correct (Loto-Quebec, 2002a).
The objective of the present study is to examine if perceived skills in sports wagering (hockey), systems or strategies are real or illusory. In order to empirically assess that point we needed to recruit sport experts who bet on hockey games. We
chose hockey bet for historical and cultural reasons. This sport is the most popular in Quebec. Nevertheless, even if hockey is a well-known sport in Quebec, can an expert bettor be considered an objective bettor ?
Gilovich (1983) demonstrated that when betting on sports, people are biased when they evaluate the information related to a game’s outcome. Success tended to be considered by bettors as obviously relevant of the skills of the winning team’s players whereas failures tended to be externally attributed (chance). Similarly, one may assume that gamblers attribute their winning wagers to their skills, and their loosing wagers to bad luck. Gilovich’s study could explain why sports bettors will continue to believe they are skilled even if their wagers are unsuccessful. So, in order to evaluate if the skills of bettors are real or illusory, it is necessary to verify if the wagers of experts sports bettors are effective. The effectiveness of wagers needs to be measured at two levels: (a) the rate of accuracy of the wagers (without considering the monetary balance of the gambler) and (b), most important, the monetary results of the wagers (i.e. do they lead to a positive money return?).
Accordingly, it was hypothesised that the accuracy of the predictions made by the participants would be superior to chance, where wagers were randomly selected. Indeed, randomness does not take into account some basic rules known by bettors (for example, based on the teams ranking, the probability of a weak team to beat a strong team is lower than the opposite). As the second hypothesis is concerned, it was hypothesised that expert bettors could achieve greater monetary gains than what can be expected from chance. In addition to these two hypothesis, we also examined, on an exploratory basis, what kind of strategies and which pieces of information participants rely on when they play Mise-O-Jeu.
Method
Participants
Thirty-five adult regular sports bettors (hockey) were recruited through media advertisement and posters in convenience stores where they could validate their wagers, in the province of Quebec. The advertisement mentioned that experts in hockey betting at the Mise-O-Jeu game were wanted for a research on gambling. Each participant received $20 for participation at the end of the experiment. All potential participants must have bet on Mise-O-Jeu game at least twice a month (36.7% played several times a week, 33.3% on a weekly basis and 30% two to three times a month).
The telephone version of the South Oaks Gambling Screen (SOGS) was used to screen for excessive gamblers (Lesieur & Blume, 1987). These were not included in the study for ethical reasons, betting being possibly harmful to them. Since
participants were French speaking, a translation of the questionnaire was used, like in Ladouceur (1991). Five participants were not included in the study for ethical reasons, since they acknowledged gambling only once a month during the experiment, even : though they said they played more often on the initial phone screening. As we defined the inclusive criterion to be at least two bets per month, it was not possible to insure that this rule would be respected by these bettors.
The final sample included in the analysis consisted of 30 male participants with a mean age of 28.6 years (SD = 9.4). The scholar level of participants was as follow: 36.7% of them had a high school degree, 36.7% of them had a college degree, and 26.7% a university degree.
Material
« Pari sportif » programs with the odds and selection slips were provided to the participants. On each selection slip, three different wagers could be placed. An
instrument containing 17 questions on hockey gambling (plus 5 demographic questions) was developed to assess the kind of information relevant to the bettors (e.g. source of documentation used when betting, importance of the precedent results of the teams and of the odds, role of home play, etc.). It contained also various
questions on how gamblers usually bet (e.g. amount of money bet, frequency of the wagers, etc.) and on their perception of the game (e.g. possibility of improving, rate of return, etc.). Answers to the questions had either to be ranked on a 0-100 scale
(beliefs on hockey betting) or on a categorical scale (yes/no or list of choices).4
Procedure
This experiment was carried out between January and April 2002, during the regular National Hockey League (NHL) season. After the initial phone assessment, participants first came at our laboratory for a 40 minutes initial meeting where they signed an informed consent, in accordance with procedures approved by our institutional ethics board. Participants who met criteria for excessive gambling were excluded and referred to appropriate resources. For 43.3% of the sample, the initial meeting took place in their community, as they were living too far away from the university. The experiment used the group format for these participants. They individually filled out the short French language Questionnaire on Mise-O-Jeu Hockey. They were also given the current “pari sportif” program and a selection slip. All the players were asked to bet as they would usually do and verbalise their thoughts, so the experimenter can understand their strategies. To ascertain that all
participants were familiarized with Mise-O-Jeu, they were told that a fixed, but fictitious amount of $2 was bet on each selection. Participants were also given instruction to fill out three wagers with the following restrictions: one selection with three games and one with four games, the last selection having no constraint. When their choice was made, they were explained in detail the second part of the study. As the participants played regularly at Mise-O-Jeu, they were invited to give their real bets. Thus, we considered that participants gambled with their own money.
The experimenter proposed 5 deadlines to participants for sending their wagers. The same instructions as in the face-to-face meeting were given for the choice of the wagers. Each participant bets on 18 selections ($2 per bet), totalling an amount of $36. The participants had to communicate their predictions either through the Internet (66.7%) or by telephone (33.3%), leaving their forecasts on an answering machine in the latter case. They were told that if they forgot to send their selections, the experimenter would communicate with them for choosing a replacement date. Before the end of the appointment, they were given a short leaflet with the
instructions, 5 selection slips, the phone number and the e-mail of the experimenter. Eventually, the experimenter proposed that they received a brief summary of the results at the end of the experiment. As for the random wagers, a random pick was made.5 Both the program’s games and the result of these games were selected with this method. Thus, in the present study the random pick served as a comparison group. The results of the participants will then be compared with this comparison group.
The accuracy of the wagers of each player was measured for the 3-games, the 4-games and the 3 to 6-games selections. The mean accuracy for all the games was
also calculated for testing our hypothesis. Accuracy is computed with the number of correct game results divided by the total number of games the participants decided to wager on. If this result is multiplied by 100, it will give the mean percentage of
accuracy obtained by gamblers. To calculate the monetary result of each participant, the total amount of money he bets ($36) was subtracted from his winning wagers. The mean of these results indicates the amount of money participants win or loose respectively.
Before validating the data given by the participants, we checked if the wagers were sent before the hockey games started. In cases they were not, participants were asked to give new wagers. Furthermore, one third of the wagers of the participants and of the random selections were checked again, to make sure that no error occurred. This reliability check was of 99.5%.
Results
Insert Table 1 here
Wagers’accuracy
Three series of forecasts of one participant were not understandable on the phone’s answering machine. As a consequence, 537 wagers were analysed (179 three-games wagers, 179 four-games wagers and 179 three to six-games wagers). The participants wagered on 1963 hockey games and the wagers amounted to $1074 in total.
An independent-sample t test was conducted to evaluate the hypothesis that hockey experts were more accurate in picking the right result for hockey games than
random selection. The test yielded significant results, revealing that the bettors (M= .473, SD = .12) were more accurate in their predictions than chance (M = .333,
SD = .01 ), f(40.87) = 5.99, p < .01. The present results are consistent with our
hypothesis. Table 1 presents the average accuracy of the wagers. These results indicate that 47.3% of the games’ prediction made by participants were correct, compared with a 33.3% accuracy obtained by the random selection.
Insert Table 2 here
Monetary results
An examination of the data indicated that the normality could not be assumed for the monetary results, Shapiro-Wilks(30) = .35, p < .01. A non-parametric Mann- Whitney U test with exact significance was conducted to evaluate if hockey experts won more money than the random selection. The result of the test, which evaluates the difference between the gamblers (M = -14.43, SD = 18.27) and the random
selection (M= -7.7, SD = 92.01), was not significant, z= -2.92, p = .99. These results revealed that the mean loss of the participants was $14.4 compared with a loss of $7.7 obtained by the random selection. The rate of return and the average amount of money made on a $2 wager bet by experts (M= 1.19) and by random selection
(M= 1.57) is displayed in Table 2. Contrary to what was expected, gamblers ($432.78 losses) did not perform any better than chance ($230.95 losses). We can see that the monetary gains of the gamblers were not in accordance with the research hypothesis.
Experts ’ strategies and ways of playing
Participants had been playing Mise-O-Jeu for 7.1 years on average (SD = 3.4). Results indicate that each time they validated their slip selections, their mean wager was $6.3 (SD = 4.2, mode = $10). They were relying on different sources of
information for betting, the most popular of them being the newspapers, TV and the Internet (Table 3).
Participants had to answer a series of questions to estimate their perception of expertise in hockey betting. All the questions had to be rated on a Likert-like scale in ten points. The scale ranged from 0 (not at all) to 100 (extremely). Participants
believed they were still able to slightly improve themselves on Mise-O-Jeu game: they answered 33.1 on average. When participants were asked to rate the degree to which their knowledge heightens their winning chances at Mise-O-Jeu hockey, they
responded 62.5 on average. Moreover, they rated high on the importance of
considering previous results of the teams when betting (75), on the question asking if home teams generally play better than the visitors (72.67) and on the question asking if high odds associated with the ties indicate they are more difficult to predict that the wins and losses (93.67). Furthermore they rated 69.17 when asked to what extent the odds influence their selections. What is more, 100% of them asserted that the odds are linked with the probabilities of gains and losses of the teams. Not
surprisingly, when asked to what extent the monetary wins are important for them when they bet, they rated 72 on average. Interestingly, when asked what the rate of return was at Mise-O-Jeu, the mean answer of the participants was 35%
(SD = 19.36). Only 21 participants are included in this mean, because 9 participants
We categorized the gamblers’ strategies to determine which was predominant. It appears that the most relied-on strategy is based on the most recent results of the teams (e.g. series of wins or losses of a team). For example, most of the participants believed that after a streak of poor performances, a “good” team was due to win.
Discussion
The aim of the present study is to evaluate whether the skills and knowledge used by gamblers are real or illusory. The results show that gamblers have an accuracy rate greater than chance when picking the results of the games. However, their monetary outcomes are not significantly higher than chance. We obtain similar results as Ladouceur et al. (1998) who analysed the wagers on horses. The results of this experiment suggest that the so-called “skills” of the sports bettors are indeed cognitive distortions. Expert bettors did not achieve better monetary gains than chance despite relying on various information that chance does not take into consideration.
The gamblers erroneously view their sport knowledge as a mean of decrypting the relative strengths of the hockey opponents. Because this did not lead to more efficacious monetary wagers, it is possible to hypothesize that these pieces of information strengthen the illusion of control. This illusion of control could even be reinforced by the accuracy gamblers achieve. In fact, the results of this study show that experts are more accurate than the random selection when picking the right results of games. However, the structure of the Mise-O-Jeu game prevents them from winning money.
It would be interesting to know how expert gamblers explain their “inefficacy” in order to avoid cognitive dissonance. Gilovich (1983) showed that gamblers maintain
the illusion of control with self deception. They consider themselves responsible and knowledgeable for their wins and they discount their losses. As Walker (1995) points out, gamblers tend to attribute successful outcomes to internal factors, whereas they attribute failures to bad luck. However, as this experiment does not inquire how participants feel about their results, the way the illusion of control may be reinforced by the near-misses cannot be stated.
Despite taking into account diverse information when betting, generally speaking hockey gamblers are not capable of wining money over a long period of time. Gamblers overestimate their ability to predict and they use unreliable
information based on past performances. This strategy resemble the “gambler's fallacy”, where chance is viewed as a self-correcting process (Clotfelter & Cook, 1993). Wood (1992) examined for baseball and basketball the predictive validity of recent performances (team streak). He comes to the conclusion that this is only a very weak predictor. Gamblers give causal explanation to what depends on chance. Moreover, Wood (1992) notes that the home field advantage also, is a weak predictor of the outcome for baseball. Even if this predictor was stronger for hockey, the
structure of the Mise-O-Jeu game is detrimental to gamblers. Players must select at least three games when betting, so the value of the already weak predictors is likely to be more diluted. At random, the probability of obtaining the correct result for a 3- games’ selection is 1:27 (1/3 x 1/3 x 1/3). For a gambler who wants to shoot the moon, the probability of right choices for a 6-games’ selection is lowered at 1:729.
The fact that participants achieved a greater accuracy than chance may help us understand their persistence at betting. The player’s perception of their accuracy may include near-misses, therefore events close to the desired outcome may
reinforce the gambling behaviour of the bettor (Delfabbro & Winefield, 1999). Contrary to games of skills, near-misses are not a useful feedback for players in a game of chance. They are not good predictors of future success when the result of the game is based on randomness (Griffiths, 1999), which is the conclusion of our study.
At a social level, the results of this study question the pertinence of lottery marketing strategies. When this experiment was carried out, the back of the page of the “Pari Sportif” programs of the Quebec’s state lottery contained the following sentence: “With Pari Sportif, Loto-Quebec offers two betting games where your knowledge of hockey could help you become a winner” (2002b). This statement is empirically unfounded and it may reinforce the gambler’s illusion of control.
Governments should be attentive to marketing campaigns developed by lottery because these may reinforce or even develop gamblers false beliefs. The gambling public should not be deluded about the nature of the game they bet on. It appears that no kind of information is valuable for winning money when betting on legal sport lotteries, at least in the case of hockey. Moreover, as the structure of the game Mise-
Ο-Jeu is slightly the same for football, baseball or soccer, it would be surprising to
find different results in these sports.
It must be emphasized that it is impossible to define objectively the concept of expertise. How can the gambler’s degree of self-qualification concerning hockey be evaluated? An expert is described as (a) “one who [...] has gained skill from
experience” or as (b) “one whose special knowledge or skill causes him to be regarded as an authority, a specialist” in the Oxford English Dictionary Online
(Simpson & Weiner, 2002). The first definition stresses the competence acquired with practice. The second one insists on the attitude of a group toward an individual. In the
latter definition, the quality of being an expert is legitimized by a community. It appears that it is problematic to screen for experts in the field of gambling, because studies indicate that gambling does not require skills or knowledge. It only relies on randomness. If there could be experts in hockey betting, it would ensue that sport betting would no longer be gambling. Furthermore, if we decide to rely on the result of the action of someone as the proof of his mastery, the problem remains unresolved. There again, how would it be possible to exclude the chance’s hypothesis?
In this study, we decided to rely on face validity (the people who pretended to be experts) and on the frequency of the participation to sports wagering. A person should at least wager twice a month to be eligible for the experiment (however, 70% of the sample played at least once a week). In fact, it is the perception of expertise that seem critical here: Our goal was to examine if the various strategies (e.g.
knowledge in hockey, considering previous teams’ results, home-ice advantage) used by participants could influence the gamblers’ monetary results. As an example,
participants reported in the Questionnaire on Mise-O-Jeu Hockey a high perception of expertise, as shown by a low rating to the question related to skill progression. This is a cue suggesting that they believe to be skilled.
Regarding the explanatory hypothesis, two points must be raised. First, as previously mentioned, all the participants in our study stated that the odds are linked with the probability associated with the different games’ outcomes. This may be a biased observation made by the participants, because the bookmakers are in fact only evaluating the gambling public’s expectation (Mallios, 2000). The bookmakers do not evaluate the real strength of the opponents; they rather make forecasts about the gambling public. This gambling public can be compared to an economic market,
where there are offer and demand for goods. In this “betting market”, there are
gamblers who “buy” odds on outcomes of games and bookmakers (the state lottery in the present study) who “sell” odds on these outcomes. To sum up, the goal of
bookmaker is to divide as evenly as possible the money bet by the public (i.e. “setting the line” 6) (Mallios, 2000). If this outcome is reached, the winners will be paid with the money of the losers. Although the Mise-o-Jeu game is slightly different from a real market, where prices are continually moving depending on demand and offer, because the odds are fixed once the program is published, the goal of the lottery is here the same as in the latter case. Keeping in mind that the line of the bookmakers is not, as Mallios (2000) points out, a forecast of a game outcome, “but rather, [their] opinion of what the public’s opinion will be” (p. 25), the odds may only indirectly represent the relative strength of the opponents, but as evaluated by the whole gambling public.
Thus, if sports gamblers faith in the odds as something representative of the chances of winning or respectively loosing of a team, they would rely on information that is not based on what they believe to be. This fact doesn’t challenge the sport knowledge of the experts, but it indicates that the experts could have been deceived to some extent by the structure of the game.
Furthermore, even if odds would have predicted well the results of games, this would not be sufficient to make money when wagering. Asch, Malkiel and Quandt (1982) have demonstrated that betting odds predict very well the order of finish in horseraces. Nevertheless, betting systematically on short, medium or long odds invariably leads to negative monetary results. Thus, it might be possible that gamblers’ tendency to link odds and results in a deterministic way (see Clotfelter &
Cook, 1993) in spite of a probabilistic one would generate or maintain an illusion of knowledge.
Being well informed on sports does not necessarily imply being well informed on the goal of oddsmaking. A lot of information is available to sports bettors; because they try to “decrypt” in this information the relative strength of the opfponents for forecasting outcomes of games, bettors could be inclined to erroneously evaluate them. Nevertheless, insufficient data was gathered with the Questionnaire on Mise-O- Jeu Hockey to precisely grasp the significance of the participants’ answers. For this reason, this discussion remains speculative. Further research should address that point and question the bettors comprehension of the nature of odds and oddsmaking. A close examination of their arguments may help to answer.
Second, concerning the rate of return put forward by the participants, the result is surprising: If the rate of return was so low (35%), who would be irrational enough to bet on this game? We formulated two hypothesis to explain this result: First of all, it could be that the gamblers don’t understand well the notion of rate of return or, it could also mean that the sample of experts are believed to be so skilled that even if the major part of the bettors loose money, they think they can make profits. This illusion of control could lead to that argument because gamblers often remember their gains and forget their losses (Wortman, 1976). Gilovitch (1983) gives another
explanation: gamblers remember well their losses but attribute them externally, without challenging their skills. So, despite their chronic bad luck, the perception of their own skills and ability to gather information lead them to believe that they are efficient. Unfortunately, the Questionnaire on Mise-O-Jeu Hockey did not include an item asking the participants to take note of their personal rate of return. A
comparison between the subjective rate of return and the rate they think the whole public can obtain would have been interesting in this study.
Interestingly, the wagers of the experts produce an average return rate
(59.5%) similar to the official approximated return rate at Mise-O-Jeu (60%) given by the state lottery (Loto-Quebec, 2001). The rate of return obtained by the random selection is 78.5%. It must be noted that the random selection includes a $456.13 win which represents 54.1% of the total amount won by chance. Indeed, we have to take into account the fact that chance can pick at random selections that produce big wins. However, we think that, on the long run, the monetary results obtained by chance and by sports gamblers would converge near the official rate of return, that is 60%,
because of the law of the great numbers. So, not only the experts do not succeed in getting sufficiently high wins to have a positive balance at the end of the study, but they also loose a significant amount of money.
As to the the ecological validity of the experiment, it must be noted that we encouraged participants who usually played together to do their wagers in a group format. One of the most interesting aspects of these participants is that they claimed to use group strategies and combinations. As some individuals prefer gambling with their friends and make use of combination in natural settings, we thought it would be valuable to offer them the opportunity to play as close as possible to a life like setting.
This study has the following practical implications. Research has already demonstrated that cognitive distortions are the core of the excessive gambling
behaviour. However, it remained unclear to what extent the knowledge of passionate sport fans were mere illusions when gambling. Clinicians could not assure their excessive sports gambling clients that knowledge and skills are not relevant to
consider when betting on hockey or football. Only scant data of this evidence was available. The results of this study, along with others (Ladouceur et al., 1998), could serve as a basis for deconstructing the myth of the “knowledgeable bettor” while doing cognitive restructuring therapy. There are no sound systems or skilled bettors, there are just many gamblers who disregard their unsuccessful outcomes, deluding themselves with an illusion that it is not a game of chance but a game of skills which they hope to acquire if they “train” well.
References
Allcock, C. (1987). An analysis of a successful racing system. In M. Walker (Ed.),
Proceedings of the Second National Conference of the National Association for Gambling Studies (1986) (pp. 179-187). Sydney, Australia: National
Association for Gambling Studies.
Asch, P., Malkiel, B. G., Quandt, R. E. (1982). Racetrack betting and informed behavior. Journal of Financial Economics, 10, 187-194.
Boulier, B. L, & Stekler, H. O. (1999). Are sports seedings good predictors?: An evaluation. International Journal of Forecasting, 15, 83-91.
Burger, J. M. (1991). The effects of desire for control in situations with
chance-determined outcomes: Gambling behavior in lotto and bingo players. Journal of
Research in Personality, 25(2), 196-204.
Clotfelter, C. T., & Cook, P. J. (1993). The “gambler’s fallacy” in lottery play.
Management Science, 39(12), 1521-1525.
Dawes, R. M. (1988). Other heuristics in evaluating probabilities. In J. Kagan (Ed.),
Rational choice in an uncertain world (pp. 90-125). Philadephia: Harcourt
Brace College Publishers.
Delfabbro, P. H., & Winefield, A. H. (1999). The danger of over-explanation in psychological research: A reply to Griffiths [Comment]. British Journal of
Psychology, 90(3), 447-450.
Gilovich, T. (1983). Biased evaluation and persistence in gambling. Journal of
Personality and Social Psychology, 44(6), 1110-1126.
Griffiths, M. (1999). The psychology of the near-miss (revisited): A comment on Delfabbro & Winefield. British Journal of Psychology, 90(3), 441-445.
Ladouceur, R. (1991). Prevalence estimates of pathological gambling in Quebec.
Canadian Journal of Psychiatry, 36(10), 732-734.
Ladouceur, R., Dubé, D., Giroux, I., Legendre, N., & Gaudet, C. (1995). Cognitive biases in gambling: American roulette and 6/49 lottery. Journal of Social
Behavior and Personality, 10(2), 473-479.
Ladouceur, R., Giroux, I., & Jacques, C. (1998). Winning on the horses: How much strategy and knowledge are needed? The Journal of Psychology, 132(2), 133- 142.
Langer, E. J. (1975). The illusion of control. Journal of Personality and Social
Psychology, 32(2), 311-328.
Lesieur, H. R., & Blume, S. B. (1987). The South Oaks Gambling Screen (SOGS): A new instrument for the identification of pathological gamblers. American
Journal of Psychiatry, 144, 1184-1188.
Loto-Quebec. (2001). Rapport annuel2000-2001 [Annual report 2000-2001]. Montreal, Quebec: Author.
Loto-Quebec. (2002a). Mise-O-Jeu regulation. Retrieved June 26, 2002, from http://192.197.135.2/p_sports/p_spr3ra.htm
Loto-Quebec. (2002b). Programme 1, Mise-O-Jeu [Leaflet]. Montreal : Author. Mallios, W. S. (2000). Modeling concepts. In The analysis of sports forecasting:
modeling parallels between sports gambling and financial markets (pp. 25-52).
Boston: Kluwer Academic.
Matsumoto, M., & Nishimura, T. (1998). Mersenne Twister: A 623-dimensionalfy equidistributed uniform pseudo-random number generator. ACM Transactions
2001, from University of Keio (Japan):
http://www.math.keio.ac.jp/~matumoto/emt.html
National Gambling Impact Study Commission. (1999). National gambling impact
study final report. Washington, DC: Author.
Numerical Technologies. (2001). Numerical Technologies Random Generator for Excel (Version 1.38) [Freeware], Retrieved November 7, 2001, from
http://www.numtech.com
Rogers, P. (1998). The cognitive psychology of lottery gambling : A theoretical review.
Journal of Gambling Studies, 74(2), 111-134.
Simpson, J. A., & Weiner, E. S. C. (Eds.). (2002). The Oxford English Dictionary
Online (2nd ed.). Oxford: Clarendon Press. Retrieved July 25, from
http://oed.com
Smith, G. J. (1992). Sucker bet on sure things: A critical analysis of sports lotteries.
Journal of Gambling Studies, 8(4), 331 -349.
Vergin, R. C., & Sosik, J. J. (1999). No place like home: An examination of the field advantage in gambling strategies in NFL football. Journal of Economics and
Business, 51, 21-31.
Walker, M. B. (1995). A sociocognitive theory of gambling involvement. In M. Argyle (Series Ed.), International Series in Experimental Social Psychology. The
psychology of gambling (pp. 130-150). New York: Pergamon Press.
Wood, G. (1992). Predicting outcomes: Sports and Stocks. Journal of Gambling
Wortman, C. B. (1976). Causal attributions and personal control. In J. H. Harvey, W. J. Ickes & R. F. Kidd (Eds.), New directions in attribution research: Vol. 1. (pp. 23-52). Hillsdale, New Jersey: Lawrence Erlbaum Associates.
York, S. (2002, March). Internet sportswagering. Paper presented at the meeting of the Gambling, law enforcement and justice system issues, Edmonton, Alberta, Canada.
Author Note
Robert Ladouceur, Michael Cantinotti and Christian Jacques, School of Psychology, Laval University, Quebec (Quebec), Canada
This research was partially supported by the “Centre Québécois d’Excellence pour la Prévention et le Traitement du Jeu” (CQEPTJ) directed by Robert Ladouceur, Ph.D. We would like to thanks Mister Serges Sévigny M.A., for his helpful comments. We also thanks the other members of the laboratory on gambling, their remarks were very useful.
Please address correspondence concerning this article to: Robert Ladouceur, Ph.D., CQEPTJ, Laval University, Quebec, Canada, G1K 7P4. Email:
Footnotes
1The Oxford English Dictionary Online (Simpson & Weiner, 2002) defines the “odds” as the “advantage conceded by one of the parties in proportion to the assumed chances in his favour; the inequality of a wager, consisting in the ratio in which the sum to be given stands to that to be received”.
2Loto-Quebec uses the terms Mise-O-Jeu and “Pari sportif” (i.e. sports wagering) also in English.
3In Canada, it is not legal to bet on the outcome of a single game. In addition to Quebec, legal sports wagers similar to Mise-O-Jeu are available in the other Canadian provinces.
4The questionnaire is available upon request.
5The program NtRand (Numerical Technologies Random Generator for Excel), version 1.38 was used (Numerical Technologies, 2001). This freeware is based on the pseudo-random generator Mersenne Twister, which was one of the most reliable free random generator available at the time of this study (Matsumoto & Nishimura, 1998). The program produced number between 0 and 1 which were further
transformed for simulating random hockey bets.
6The line is the odds quoted by a bookmaker, and also, the point-spread predicted in a football game, from which such odds are calculated (Simpson & Weiner, 2002).
Table 1
Average accuracy of bets
Kind of bet Expert selection Random selection
on 3-games 48.6% 33.9%
on 4-games 49.3% 35.6%
on 3 to 6-games 44.2% 30.8%
Table 2
Average amount of money ($) returned on bets wagered by experts and by random selection
Kind of bet Expert selection Random selection
$2 on 3 games 1.69 0.69
$2 on 4 games 1.25 3.67
$2 on 3 to 6 games 0.64 0.36
Table 3
Sources of information utilized by participants
Source of information Percentage of participants utilizing it
Newspapers 80%
Television 63%
Internet 43%
Radio 37%
Advice of friends 27%
Computer or pocket calculator 20%
Advice of “experts” 10%
Other 10%
Bien qu’il soit parmi les jeux d’argent qui génèrent les plus grosses mises d’ensemble (illégales pour la plupart) aux Etats-Unis (National Gambling Impact Study Commission, 1999), le pari sportif a été peu étudié. Pourtant, le fait de considérer le pari sportif comme un jeu de hasard peut sembler contre-intuitif. En effet, les résultats sportifs sont généralement considérés comme découlant, du moins en partie, des capacités sportives des athlètes. Le gens considèrent qu’au sport, les compétiteurs ont un degré de contrôle relativement important sur leurs résultats (Keren & Wagenaar, 1988). Néanmoins, est-il légitime de conclure que la
connaissance d’un certain nombre de paramètres réputés influencer les résultats augmente les chances de prédire l’issue correcte d’un événement sportif ? Autrement dit, les connaissances et les habiletés sur lesquelles certains parieurs se basent sont- elles réellement utiles ou, à l’inverse, sont-elles le reflet d’une distorsion cognitive ?
La présente étude avait pour objectif principal de vérifier si les résultats monétaires obtenus par des joueurs experts de pari sportif sur le hockey s’avèrent supérieurs à ceux qui découlent d’une sélection aléatoire. De même, il a été vérifié si le taux de précision atteint par les parieurs, mesuré par !’intermédiaire du nombre de résultats corrects pronostiqués, était lui aussi plus élevé que celui généré par le hasard. Enfin, les différentes informations considérées pertinentes par les parieurs pour effectuer les mises ont été relevées sur une base exploratoire.
L’analyse des résultats démontre que l’hypothèse d’un retour monétaire des parieurs experts supérieur à celui obtenu par le hasard ne peut être soutenue. Au niveau de la précision des parties prises une à une, les parieurs obtiennent toutefois un taux significativement plus élevé que celui atteint par le hasard. Quant aux
informations sur lesquelles les participants se basent, celles-ci sont diversifiées et font appel à un grand nombre de connaissances sur les équipes.
Cette étude empirique sur les résultats de parieurs experts indique donc que les connaissances et le jugement sportifs d’amateurs de hockey ne suffisent pas pour leur permettre d’atteindre des gains monétaires positifs. Malgré un taux de précision supérieur à celui obtenu par la sélection aléatoire, les participants ne parviennent pas à “battre le hasard“. Au regard de ces résultats, il est possible de supposer que le fait de s’approcher des résultats corrects des parties constitue un ”presque gain”, tel que définit par Reid (1986), qui pourrait expliquer la persistance à jouer malgré des
résultats négatifs (Gilovich, 1983). En outre, les résultats de cette étude s’ajoutent à ceux obtenus par Ladouceur et al. (1998), ainsi que Wood (1992) et indiquent que les habiletés des parieurs jouent un rôle minime dans le cadre des paris sur événements sportifs.
Les résultats de cette recherche soutiennent donc l’hypothèse qui affirme que les jeux de pari sur événements sportifs sont semblables aux autres jeux de hasard et d’argent. Conséquemment, il semble judicieux de considérer que les différentes habiletés et connaissances dont les parieurs mentionnent faire usage sont une manifestation de l’illusion de contrôle.
Allcock, C. (1987). An analysis of a successful racing system. Dans M. Walker (Éd.),
Proceedings of the Second National Conference of the National Association for Gambling Studies (1986) (pp. 179-187). Sydney, Australia: National
Association for Gambling Studies.
Asch, P., Malkiel, B. G., Quandt, R. E. (1982). Racetrack betting and informed
behavior. Journal of Financial Economics, 10, 187-194.
Boulier, B. L, & Stekler, H. O. (1999). Are sports seedings good predictors?: An evaluation. International Journal of Forecasting, 15, 83-91.
Burger, J. M. (1991). The effects of desire for control in situations with
chance-determined outcomes: Gambling behavior in lotto and bingo players. Journal of
Research in Personality, 25(2), 196-204.
Clotfelter, C. T., & Cook, P. J. (1993). The “gambler’s fallacy” in lottery play.
Management Science, 39(12), 1521-1525.
Dawes, R. M. (1988). Other heuristics in evaluating probabilities. Dans J. Kagan
(Éd.), Rational choice in an uncertain world(pp. 90-125). Philadephia: Harcourt Brace College Publishers.
Delfabbro, P. H., & Winefield, A. H. (1999). The danger of over-explanation in
psychological research: A reply to Griffiths [Commentaire]. British Journal of
Psychology, 90(3), 447-450.
Department for Culture, Media and Sport. (2001). Gambling review report. London: The Stationery Office.
Eurispes. (2000). Giochi, scommesse e lotterle: Italian¡ d’azzardo. Indagine sul gioco
in Italia. Extrait le 6 août 2002, sur le site de l’Institut européen d’étude
politique, économique et social
http://www.eurisped.com/Eurispes/azzardo/default
Gilovich, T. (1983). Biased evaluation and persistence in gambling. Journal of
Personality and Social Psychology, 44(6), 1110-1126.
Griffiths, M. (1999). The psychology of the near-miss (revisited): A comment on Delfabbro & Winefield [Commentaire]. British Journal of Psychology, 90(3), 441-445.
Keren, G., & Wagenaar, W. A. (1988). Chance and skill in gambling: A search for distinctive features. Social Behaviour, 3, 199-217.
Ladouceur, R. (1991). Prevalence estimates of pathological gambling in Quebec.
Canadian Journal of Psychiatry, 36(10), 732-734.
Ladouceur, R., Dubé, D., & Bujold, A. (1994). Gambling among primary school students. Journal of Gambling Studies, 10(4), 363-370.
Ladouceur, R., Dubé, D., Giroux, I., Legendre, N., & Gaudet, C. (1995). Cognitive biases in gambling: American roulette and 6/49 lottery. Journal of Social
Behavior and Personality, 10(2), 473-479.
Ladouceur, R,, Giroux, I., & Jacques, C. (1998). Winning on the horses: How much strategy and knowledge are needed? The Journal of Psychology, 132(2), 133- 142.
Langer, E. J. (1975). The illusion of control. Journal of Personality and Social
Lesieur, H. R., & Blume, S. B. (1987). The South Oaks Gambling Screen (SOGS): A new instrument for the identification of pathological gamblers. American
Journal of Psychiatry, 144, 1184-1188.
Loto-Québec. (2001). Rapport annuel2000-2001. Montréal, Québec: Auteur. Loto-Québec. (2002a). Mise-O-Jeu regulation. Extrait le 26 juin 2002, sur
http .7/192.197.135.2/p_sports/p_spr3ra.htm
Loto-Québec. (2002b). Programme 1, Mise-O-Jeu [Brochure]. Montréal : Auteur. Mallios, W. S. (2000). Modeling concepts. Dans The analysis of sports forecasting:
modeling parallels between sports gambling and financial markets (pp. 25-52).
Boston: Kluwer Academic.
Matsu moto, M., & Nishimura, T. (1998). Mersenne Twister: A 623-dimensionally equidistributed uniform pseudo-random number generator. ACM Transactions
on Modeling and Computer Simulation, 8(1), 3-30. Extrait le 7 novembre 2001,
sur le site de l’Université de Keio (Japon)
http://www.math.keio.ac.jp/~matumoto/emt.html
Ministère de la Justice du Canada. (2001). Partie VII. Maisons de désordre, jeux et paris. Dans Code criminel (Articles 197 à 213). Extrait le 25 août 2002, sur http://lois.justice.gc.ca/fr/C-46/17713.html
National Gambling Impact Study Commission. (1999). National gambling impact
study final report. Washington, DC: Auteur.
Numerical Technologies. (2001). Numerical Technologies Random Generator for Excel (Version 1.38) [Programme informatique, Gratuiticiel]. Extrait le 7 novembre 2001, sur http://www.numtech.com
Reid, R. L. (1986). The psychology of the near-miss. Journal of Gambling Behavior, 2(1), 32-39.
Rogers, P. (1998). The cognitive psychology of lottery gambling : A theoretical review.
Journal of Gambling Studies, 14(2), 111-134.
Sénat français. (2002). Les jeux de hasard et d’argent en France. Rapport
d’information. Extrait le 6 août 2002, sur
http://www.senat.fr/rap/r01 -223ΛΌ1 -2239.html
Simpson, J. A., & Weiner, E. S. C. (Éds.). (2002). The Oxford English Dictionary
Online (2ème éd.). Oxford: Clarendon Press. Extrait le 25 juillet 2002, sur http://oed.com
Smith, G. J. (1992). Sucker bet on sure things: A critical analysis of sports lotteries.
Journal of Gambling Studies, 8(4), 331-349.
Société Québécoise d’information Juridique. (1983). Article 1390, General Attorney, of Canada vs. Loto-Québec. Dans Annuaire de Jurisprudence du Québec (p. 311). Montréal : Auteur.
U.S. Senate Committee on Commerce, Science, and Transportation. (2001, Mai).
Committee approves bill to stop betting on amateur sports. Extrait le 23 juillet
2002, sur http://www.senate.gov/~commerce/press/107-47.html
Vergin, R. C., & Sosik, J. J. (1999). No place like home: An examination of the field advantage in gambling strategies in NFL football. Journal of Economics and
Walker, Μ. B. (1995). A sociocognitive theory of gambling involvement. Dans Μ. Argyle (Éd. série), International Seríes in Experimental Social Psychology. The
psychology of gambling (pp. 130-150). New York: Pergamon Press.
Wood, G. (1992). Predicting outcomes: Sports and Stocks. Journal of Gambling
Studies, 8(2), 201-222.
Wortman, C. B. (1976). Causal attributions and personal control. Dans J. H. Harvey, W. J. Ickes & R. F. Kidd (Éds.), New directions in attribution research: Vol. 1. (pp. 2352־). Hillsdale, New Jersey: Lawrence Erlbaum Associates.
York, S. (2002, Mars). Internet sportswagering. Contribution présentée à la
conférence Gambling, law enforcement and justice system issues, Edmonton Alberta, Canada.
Centre Québécois d’excellence pour la prévention et le traitement du
leu
UNIVERSITÉ
LAVAL
Formulaire de consentement
Iitredu.p_rojet: PERCEPTION DU HASARD ET GAINS SELON LA RÉGULARITÉ DU JEU CHEZ DES JOUEURS DE PARIS SPORTIFS SUR LE HOCKEY
N° d’approbation du comité : 2001-186 a1
1. Le but de l'étude consiste d'une part, à vérifier si la perception du hasard chez des amateurs de paris sportif (hockey) diffère selon leur fréquence de jeu et si ces derniers perçoivent de façon différenciée un jeu de loterie active traditionnel (6/49, par exemple) par rapport à un jeu de paris sportifs. D’autre part, cette recherche veut déterminer quelle est l’efficacité des mises des parieurs.
2. La recherche à laquelle je désire participer consiste en trois phases : (a) une première entrevue individuelle durant laquelle j’effectue des pronostics et je réponds à un questionnaire (environ 30 minutes) et (b) cinq brefs appels téléphoniques (environ 3 minutes chacun) ou 5 courriers électroniques (selon ma préférence) au cours desquels je donnerai mes pronostics pour le programme de pari sportif courant. Une fois la dernière communication effectuée, !’expérimentateur communiquera avec moi pour répondre à mes éventuelles questions. Enfin, (c) je devrai remplir un questionnaire qui me sera envoyé par voie postale. Je retournerai celui-ci à l’aide d’une enveloppe pré-affranchie par !’expérimentateur.
3. Je sais que suite à la séance de jeu et lors des entretiens téléphoniques ou de l’envoi des courriers électroniques, !’expérimentateur conservera une copie de mes pronostics afin d’évaluer la justesse de mes prédictions. Je consens librement à ce que les informations que je fournirai par mes pronostics et par les questionnaires soient utilisées par les responsables de la recherche sans que mon identité soit divulguée.
4. En plus de contribuer à l’avancement des connaissances liées à la prévention et au traitement du jeu pathologique, ma participation me donne l’occasion de réfléchir sur mes habitudes de jeu et sur ma perception des jeux de hasard et d’argent. Le désavantage de ma participation est le temps que je dois y consacrer et possiblement, un certain stress à devoir effectuer les expériences susmentionnées dans un laboratoire en compagnie d’un chercheur.
