The study of economics is based on key concepts such as incentives, efficiency, marginality and tradeoffs. Economic research has hypothesized and tested for how economic agents behave after taking each of these into account. In order for agents to meet their objectives it is sometimes the case that they intentionally keep their behaviors out of sight. However, economic theory can be used to search for patterns of observed behaviors from which the unobserved behaviors can be inferred. This dissertation performs this kind of analysis by observing the behavior of sports participants. Chapter 1 is an application of Becker's (1968) economic model of crime by using an econometric model to search for the presence of National Basketball Association (NBA) referees who bet on NBA games. The placement of these bets is not observed since a referee who bets on a game does so illegally and therefore hides his betting activity to prevent detection. A referee who places a bet on a game he also officiates has an incentive to manipulate to improve his chances of winning the bet. At the same time he should also be mindful to manipulate in a way that lowers his chances of being detected. The referee's observed behaviors through detailed play-by-play data are used to look for patterns hypothesized to be consistent with manipulation. The results suggest that former NBA referee Tim Donaghy, who was found to have bet on NBA games, did behave in ways consistent with manipulation. One other referee also appears to engage in the same type of behavior but stops once Donaghy is detected. Chapter 2 is an application of Fama's (1970) Efficient Market Hypothesis (EMH). Typically, the EMH is tested in the financial markets but some research tests for it in the sports betting markets so that the question becomes whether or not the betting market odds fully reflect all of the available relevant information. This chapter tests to see how completely National Football League (NFL) bettors use information called the circadian advantage. This occurs when a game is played in the evening, Eastern Time, between teams that are based on opposite coasts and always favors the better rested West Coast team. A regression model designed to test for market efficiency finds that the advantage is not fully reflected in the odds so that bets on the West Coast team are underpriced. In a majority of games that involve a circadian advantage most of the money is wagered on the overpriced East Coast team. A conclusion that ties these results together is that the bookmakers restrict the amount bet from informed bettors who tend to win their bets and who are aware of the circadian advantage, and adjust the odds just enough to bait uninformed bettors who are unaware of the circadian advantage into placing wagers on the team that is overpriced. Given these dynamics, it is the bookmakers who profit from the information contained in the circadian advantage. Chapter 3 revisits the NFL betting market but instead estimates the extent to which bettors place wagers based on sentiment for a team that is unrelated to relevant measures of relative performance along the lines of speculative investment outlined by Graham and Dodd in 1934 (2009). The results show that more bets tend to be placed on teams for which bettors have high sentiment and fewer bets are placed on teams for which bettors have low sentiment. However, the market odds appear to be using sentiment unbiasedly, leading to the conclusion that contrarian bettors place wagers opposite the sentimental bettors. While the market as a whole is efficient in the use of sentiment, losers tend to be bettors who wager with sentiment and winners tend to be bettors who wager against sentiment.

This thesis deals with the impacts of amendments of Act on Lotteries and Other Like Games effective since 2012 on sports betting operators, state budget of the Czech Republic and funding of Czech sport. Furthermore, this paper analyses the issue of cross-border provision of gambling services in the EU and examines the compliance of Czech regulatory framework with EU law. Finally, it proposes solutions to identified weaknesses.

01 Abstract Betting on sporting events can be perceived by the general public as a game of chance. In the professional literature, however, betting on football matches is treated in the same way as other financial markets, where in the event of a violation of the theory of efficient markets due to the occurrence of inefficiency, there are opportunities for investors to obtain abnormal returns. The main goal of this work is to create a model capable of predicting the results of football matches on the basis of historical data better than bookmakers are able to do and test the effectiveness of the Czech betting market for football matches of the Czech highest football league. The first part of the thesis contains a more detailed presentation of the theory of efficient markets, a comparison of financial and betting markets and sources of possible inefficiency in betting markets. The second and third parts present data, models and their possible modification to increase the accuracy of estimates. The fourth part describes the results of testing individual models and subsequent simulations of betting strategies. The fifth part contains a conclusion and discussion of the results, including an indication of possible alternatives to follow-up research. The results of simulations of betting strategies confirm...

This study aimed to compare the accuracy of machine learning models with the probabilities generatedby sports betting companies in predicting the outcome of football matches. The study also investigatedthe impact of different feature combinations on the performance of machine learning models for predicting football match outcomes. The study used data from various sources of the Swedish football leaguebetween the seasons 2018-2022. The comparison between the model’s predictions and the probabilitiesgenerated by sports betting companies showed that the model’s predictions were more accurate. SupportVector Machines(SVM) performed the best with an accuracy of 52.4 percent compared to the bettingcompanies at 40.4 percent. The results also showed that different feature combinations can have a significant impact on the performance of machine learning models for predicting football match outcomes butthe importance of features varied depending on the selection method used. The study used four different feature selection approaches: filter methods, Lasso, Ridge, and PCA, to identify the most importantfeatures for prediction. Overall, the results of this study suggest that machine learning models can outperform sports bettingcompanies in predicting football match outcomes and that the choice of feature combination can have asignificant impact on model performance. Further research is needed to explore these findings in moredetail and to investigate the usefulness of different feature selection techniques at different points in theseason.

In 2018, the United States Supreme Court overturned PASPA, a law which had previously deemed sports betting illegal. Following this ruling, states have already or have begun passing legislation which legalizes sport betting. As legalization continues to sweep the nation, an untapped domain of research has emerged. From a sport management perspective, there is a new, highly lucrative sport industry with which there is minimal research. The main purpose of this research project is to examine how bet presentation influences consumer behavior related to sports gambling. Specifically, the role that bet format presentation has on consumers’ willingness to bet and the amount they are willing to bet. Additionally, the potentially mediating effects of numeracy and team identification were examined. Participants (N=703) were recruited from the United States, United Kingdom, and Australia, as these locations natively use different forms of bet presentation (American, fractional, and decimal). This study utilized a Latin square experimental design that examined whether participants were willing to bet more money when shown American odds first compared to fractional odds first. Further, evidence was provided demonstrating the positive mediating influence of team identification, and the influence of subjective numeracy. Practically, the results from this study can inform sports betting organizations, sports betting consumers, as well as government and industry regulators. Theoretically, knowledge is contributed to the domains of sport management, behavioral pricing, and appraisal theory literature. / Tourism and Sport

Since the U.S. Supreme Court lifted the federal ban on sports gambling, the popularity of sports gambling continues to increase. This has left the sport industry, including academics, interested in examining the drivers of sports gambling participation and their influence on consumers’ betting behavior. This dissertation includes three essays considering motivations to engage in sports gambling. While all focus on sports gambling, each of these three standalone essays embrace a different focus to explore sports gambling motivations and betting behavior. First, Essay One explores the differences in motivation and perception of skill versus luck between daily fantasy sports (DFS) and sports betting participants. Next, Essay Two investigates the interplay between motivations and game characteristics on betting intentions. Finally, Essay Three explores the effects of different marketing promotions and their fit with consumers’ regulatory focus on consumers’ betting behavior. Collectively, this research will provide insights and understandings of different drivers of sports gambling and their influence on consumer behavior regarding sports gambling. / Tourism and Sport

This paper presents a study on predicting football match outcomes using various machine learning models, aiming to outperform traditional bookmakers. The primary objectives were to develop predictive models, identify key factors influencing match results, and assess the potential monetary value of the models. The study utilized an ANOVA test for feature selection, revealing that differences in team quality metrics, such as Elo ratings and FIFA player ratings, are among the most significant predictors. Despite achieving competitive accuracy, with Linear Discriminant Analysis (LDA) reaching 68.8 %, the models generally underperformed compared to bookmakers' odds, which also achieved 68.8 % accuracy. Betting strategies were tested over 109 matches, where the probability betting strategy yielded profits for all models, with LDA achieving a 21.5 % profit, slightly surpassing the bookmakers' strategy. However, the expected value betting strategy resulted in losses, indicating a challenge in predicting non-favorite outcomes. The findings suggest that while the machine learning models developed in this project show promise in predicting match results, they require further refinement to consistently outperform traditional bookmakers, especially in identifying valuable betting opportunities.

Cette thèse a pour but de présenter le marché des paris sportifs plus précisément de montrer en quoi ce dernier constitue un cadre d'observations simplifié suffisamment proche des marchés boursiers pour tester la théorie de l'efficience informationnelle et aboutir à des conclusions unanimes concernant sa validité empirique. En premier lieu, nous concentrons notre attention sur la forme faible de l'efficience informationnelle et plus précisément sur une anomalie connue sous le nom du Favourite Longshot Bias, qui a été recensée aussi bien dans le cadre des paris sportifs que celui des marchés boursiers. A l'aide d’un vaste échantillon de données, nous démontrons que les coûts de transaction et les préférences des parieurs ont un impact significatif sur le niveau des cotes proposées par les bookmakers et donc sur la structure des prix. Par ailleurs, nous discutons de la rationalité des parieurs et nous montrons en quoi le comportement des parieurs n’est pas si différent de celui des investisseurs sur les marchés boursiers. En second lieu, nous analysons en détails la forme forte de l'efficience informationnelle et plus précisément la pertinence de la fourchette en tant qu'indicateur de délits d'initiés dans le cadre des paris sportifs. / The aim of this thesis is to introduce the sports betting market, and more precisely to show how this market forms a simplified framework for observations, close enough to stock markets to test the informational efficiency theory, and lead to unanimous conclusions about its empirical validity. Firstly, we focus on the weak form of informational efficiency and more particularly on an anomaly known as the Favourite Longshot Bias, which was listed in sports betting markets, as well as in stock markets. Thanks to a vast data sample, we prove that transaction costs and bettors’ preferences have a significant impact on odds fixed by bookmakers, and consequently on prices’ structure. Moreover, we also discuss bettors’ rationality and we show how bettors’ behaviour is not so different from that of investors in the stock markets. Secondly, we provide a detailed analysis of the strong form of informational efficiency and more precisely of the spread as an indicator of insider trading in the sports betting market.

Online sports betting is an industry that has seen a huge surge in popularity in the US since 2018, and the companies behind this surge have enjoyed exponential increases in profits. Accompanying the increased interest in and profits of online sports gambling is a large spike in advertisements created by companies that offer online sports betting services. Although gambling is an activity that is recognized as highly addictive and potentially destructive for certain individuals, the companies that offer sports gambling services in the US face little to no government regulation in how they advertise their services. This study aimed to investigate the advertising narratives deployed by the four most popular online sports betting companies in the US (FanDuel, DraftKings, Caesars Sportsbook, and BetMGM) in terms of their frequency, using the precision afforded by the method of qualitative content analysis. Additionally, this study deployed semiotic analysis with the constructionist approach to representation and the concept of myth to examine how the most frequently used narratives in online sports betting advertisements draw on broader systems of meaning in order to persuade audiences. This study found that FanDuel, DraftKings, Caesars Sportsbook, and BetMGM most frequently used narratives of online betting and celebrity feature in their advertisements. Using semiotic analysis, this study found that the narratives of online betting and celebrity feature often translated to mythologized and often unrealistic depictions of sports betting which represented it as an activity that allowed sports bettors to align themselves on the same hierarchical level as professional athletes and Hollywood celebrities, while largely ignoring the risks that are inherently attached to gambling. The findings of this study stand as an exploratory accounting and examination of the currently unregulated advertising practices of the largest sports gambling companies in the US.

In this work, we propose a regressional framework, built on the work ”Deep Order Flow Imbalance: Extracting Alpha at Multiple Horizons from the Limit Order Book” by Kolm, et al. (2023), for predicting short term returns of odds on binary betting exchange markets. Using the framework, we apply five different deep learning models that leverage order book data from tennis betting exchanges during the calendar month of July 2023 with the purpose of examining the predictive capabilities of deep learning models in this setting. We train each model on either raw limit order book states or order flow. The models predict the returns of the best available odds returns on five different short term time horizons on the four order book sides, back and lay for each of the two players in a given tennis match. Applying windowing, for each vector prediction we use the 100 latest market messages consisting of 81 features (odds and volumes per the ten first levels in the order book and time delta between market messages) in the case of the raw limit order book state and 41 features (order book flow per the ten first levels in the order book and time delta between market messages) in the case of the order book flow. All code is written in Python and run on Google Colab, leveraging cloud computing, off-the-shelf models and popular libraries, TensorFlow and Keras, for data processing and pipelining, model implementation, training and testing. The models are evaluated relative to a benchmark in the form of a naive predictor based on the average odds returns on the training set. The models do not converge towards an optimal parameter composition duringtraining, indicating low predictive capabilities of the input data. Despite this, we generally find all models to outperform the benchmark on the lay order book sides and while some perform better than others, we see similar relative performance distributions within each model across horizon-order book side combinations. To enhance discussion and suggest the direction of future research we examine relationships between key game characteristics such asthe variation of odds returns and the accuracy of predictions on a given market.