Science of Sports Betting

Understanding sports betting markets through reverse engineering of market lines, alternative lines, and systematic market analysis; an NBA alternative-lines case study; empirical distributions, statistical assumptions, random variables, expected value, edge, variance, Gaussian methods, exploratory data analysis, and an introduction to NHL historical odds work in R.

Different approaches to profitable sports betting, including top-down, bottom-up, and hybrid workflows; inter-book mispricings, line grinding, market coherence, steam chasing, and methods of evaluation such as naive backtesting, coherent backtesting, shrinkage, closing line value, z-scores, goodness-of-fit testing, and practical caveats.

Regression, residuals, market resistance, market support, parimutuel versus fixed odds, fractional odds, and the definition and empirical justification of closing line value, with real-data demonstrations showing how CLV and expected value interact in practice.

Closing line value for spreads and totals, derivation and use of CLV calculations, alternative-line approaches, partial markets, teasers, historical empirical and Gaussian approaches, and the theory and practical execution of middling across NBA, NFL, and NHL markets.

Record keeping, additive versus multiplicative dynamics, growth-rate maximization, loss aversion, prospect theory, overbetting, the Kelly criterion, the components of an edge, and the relationship between arbitrage, Kelly sizing, and real-time hedging.

A guest session with "JM," a renowned sports bettor and market maker leading a dynamic licensed Maltese syndicate focused on U.S. sports trading, offering perspective on how advanced betting ideas operate in live professional environments.

A review of the top-down approach before introducing bottom-up and hybrid modeling, including data acquisition and quality control, file management, variables, distributions, logistic regression, deep learning, model life cycle, backtests, and the transition from exploratory work to structured model development.

Best practices and common pitfalls in model evaluation, including coherence checks, market checks, residual analysis, flat versus Kelly betting, in-sample and out-of-sample testing, variable development, weighting schemes, and the practical differences between backtesting and live betting.

Shrinkage from both theoretical and applied perspectives: calibration, model refinement, loss functions, regression shrinkage, evaluation metrics, variable selection, model assumptions, and how shrinkage helps refine probabilities, totals, and live-betting execution.

A deeper study of optimal betting and Kelly sizing, including fractional Kelly, multiple-outcome Kelly, parlays, correlated outcomes, growth-rate comparisons, expected value under different correlation structures, and practical sizing considerations in sportsbook and parimutuel settings.

Modeling beyond Gaussian assumptions, including binary outcomes, logistic regression, shrinkage for probabilities and moneylines, player and team relative-ability models, binomial and Poisson frameworks, generalized linear models, hybrid modeling, and robustness checks.

Four in-depth case studies showing real applications of course concepts in profitable betting, with examples drawn from NFL, NBA, NHL, and MLB markets and covering hybrid modeling, uncertainty pricing, distributional assumptions, and reverse engineering of model parameters.

A guest lecture by Dr. William T. Ziemba, Alumni Professor (Emeritus) of Financial Modeling and Stochastic Optimization at the Sauder School of Business, University of British Columbia, and the author of several of the most influential books on gambling and betting theory.