Acknowledgments
A book of this scope depends on the work of many individuals and communities, and we owe debts of gratitude to those whose published and openly shared contributions made it possible.
Our first thanks go to the academic researchers whose work forms the intellectual foundation of this textbook. The scholars who have studied sports betting markets with the same rigor applied to financial markets—publishing in journals of statistics, econometrics, operations research, and computer science—have elevated the field from folklore to science. We are particularly indebted to the researchers in market microstructure whose frameworks for understanding price formation, information asymmetry, and market efficiency translated so naturally to the study of betting lines. The growing body of peer-reviewed literature on prediction markets, rating systems, and machine learning for sports forecasting gave us a solid base on which to build. We have cited these contributions throughout the text, but a citation can only partially express the intellectual debt we owe.
We are equally grateful to the sports analytics community at large. The annual MIT Sloan Sports Analytics Conference and similar gatherings have created a culture of openness and cross-pollination that benefits everyone in the field. The practitioners who share analyses on blogs, in podcasts, on social media, and in online forums—often at the cost of eroding their own competitive advantage—have created a public knowledge base of extraordinary depth. The public discussion in sports analytics communities on social media, Discord servers, and specialized forums is a steady source of ideas, critiques, and reality checks.
The open-source software community deserves special recognition. This book's code is built entirely on freely available tools: Python and its scientific ecosystem, including NumPy, pandas, scikit-learn, statsmodels, PyTorch, XGBoost, and many others. The developers and maintainers of these projects—most of them volunteers or modestly funded—have democratized quantitative work in ways that would have been unimaginable a generation ago. We also acknowledge the creators of sports data APIs and open datasets who have made the raw material of sports analytics accessible to independent researchers and hobbyists. Without open data, the exercises and case studies in this book could not exist in their current form.
A note on how this book was made: it was drafted with AI assistance and is published by DataField.Dev. It has not been reviewed by beta readers, an editorial team, or subject-matter experts, so any errors or shortcomings are entirely the book's own. That makes our readers especially important. If you are an instructor, practitioner, student, or beginner and you find an error, an unclear explanation, a flawed assumption, or a chapter that fails to deliver on its promise, please tell us—bluntly, if need be—through datafield.dev/about.html. Corrections and suggestions will make future editions better.
Finally, we acknowledge the inherent tension in writing a textbook about a competitive domain. Every insight we formalize and publish is, in some sense, an edge shared and therefore diminished. We have proceeded with the conviction that the long-term health of the field depends on raising the general level of understanding, that education is not a zero-sum game, and that the problems ahead are interesting enough to sustain progress for those willing to do the work. We hope our readers agree.