Date of Submission
Campus Only Senior Thesis
Bachelor of Arts
Sports betting has been a fascinating topic that piques the interest of many mathematicians and statisticians over the centuries. Nowadays, equipped with access to massive record data, similar attempts have taken on much more data-driven approaches. This thesis devises a strategy for in-game sports betting, specifically the National Basketball Association (NBA) basketball games. We first explain the basic rules of sports betting and describe the data set used for this thesis. Then, we apply our method to the data set. Our strategy consists of two component, updating on teams' strength and allocating bets through the Kelly betting system. Finally, we test our strategy across different seasons. The entire project is conducted through the usage of Python in Jupyter Notebook.
Overall, the strategy devised in this paper still incurs a loss through its betting over a five-year time span, but it is able to generate significant returns in 2 of the 5 seasons, with one of them achieving a 45\% return. This finding suggests that while there are inefficiencies to exploit in short term, the bookmakers that set the odds still have a much greater advantage than participants.
Despite the bookmaker's advantageous positioning, this thesis also delves into future possible improvement of the strategy. The model built in this thesis should serve as a starting point to build on. The approach we employ here also provides a general framework to approach similar kinds of problems.
Song, Yuan, "Sports Betting: NBA Games as a Poisson Point Process" (2020). CMC Senior Theses. 2537.
This thesis is restricted to the Claremont Colleges current faculty, students, and staff.