Mark Glickman, Senior Lecturer on Statistics, Department of Statistics, Harvard University

Title: Rating competitors in games with strength-dependent tie probabilities

Abstract: Competitor rating systems for head-to-head games are typically used to measure playing strength from game outcomes. Ratings computed from these systems are often used to select top competitors for elite events, for pairing players of similar strength in online gaming, and for players to track their own strength over time. Most implemented rating systems assume only win/loss outcomes, and treat occurrences of ties as the equivalent to half a win and half a loss. However, in games such as chess, the probability of a tie (draw) is demonstrably higher for stronger players than for weaker players, so that rating systems ignoring this aspect of game results may produce strength estimates that are unreliable. We develop a new rating system for head-to-head games that explicitly acknowledges a tie as a third outcome, and that the probability of a tie may depend on the strengths of the competitors. Our approach relies on time-varying game outcomes following a Bayesian dynamic modeling framework, and that posterior updates within a time period are approximated by one iteration of Newton-Raphson evaluated at the prior mean. The approach is demonstrated on a large dataset of chess games played in International Correspondence Chess Federation tournaments.

Mark Glickman is a Senior Lecturer on Statistics at Harvard University and a recognized leader in sports analytics and statistical ranking methodology. He is best known for developing the Glicko and Glicko-2 rating systems, which extend the Elo framework by explicitly modeling uncertainty in player strength and have been adopted in chess, esports, and other competitive domains. His research spans Bayesian inference, hierarchical modeling, paired-comparison models, and decision-making under uncertainty, with applications in sports, games, public policy, and the social sciences. Dr. Glickman has played a visible leadership role in the statistics and data science community through professional society service, conference organization, and scholarly leadership. Particularly relevant to CSAS, he has been one of the organizers of the New England Symposium on Statistics in Sports (NESSIS), a long-running and highly influential forum that brings together students, researchers, and practitioners working at the interface of statistics, data science, and sports analytics. His work is widely cited for its clarity, practical relevance, and principled statistical foundations, and it has shaped how modern performance evaluation problems are formulated and solved.

Dean Oliver, Sports Analytics Pioneer, Project Specialist, Sports Statistics ESPN

Title: Why division of (micro)credit helps in sports

Abstract: At ESPN, we have a wide receiver rating, a pass blocker rating, a quarterback rating, a net points rating for various basketball leagues. A player rating is useful because everyone talks about how good (or bad) players are --- fans, analysts, coaches, medical staff --- and just about everyone does it based on their eyes. Or, frankly, based on what the media feeds them. But I worked for NBA teams and you know what? Subjective opinions aren't very useful there. You gotta have something that does something that good scouts can't match --- you have to have something that sees all the plays. You gotta have something that not only says who is good, but why, when, and how. To do that needs metrics that work at a very granular level. And you know what you need to do that? A way to divide credit at a very fine level. I'll introduce some of the math and some of the intuition behind doing this the right way.

Dean Oliver revolutionized basketball analytics with his book, Basketball on Paper. He brought unique insight to the game through a clear statistical breakdown of how the game worked. That led him to work in both management and coaching staffs in the NBA, as well as to help found ESPN Analytics, where he is now. His work covers player value, game strategy, sports psychology, and the statistical tools to investigate all of them in ways no one else has. His follow-up book, Basketball beyond Paper, documents how basketball analytics is evolving in a more tech-driven industry, as well as the challenges he faced in growing it.

Panel Discussion: TBA

Abstract: TBA

Dean Oliver revolutionized basketball analytics with his book, Basketball on Paper. He brought unique insight to the game through a clear statistical breakdown of how the game worked. That led him to work in both management and coaching staffs in the NBA, as well as to help found ESPN Analytics, where he is now. His work covers player value, game strategy, sports psychology, and the statistical tools to investigate all of them in ways no one else has. His follow-up book, Basketball beyond Paper, documents how basketball analytics is evolving in a more tech-driven industry, as well as the challenges he faced in growing it.

Alok Pattani is a Data Science Advocate at Google, where he shows how to use Google Cloud tools for data science, in sports and otherwise. He is a sports analytics expert, and has played a significant role in Google Cloud’s sports-related partnerships, including basketball data-related analysis used during Google’s NCAA Tournament campaigns. Alok joined Google Cloud after 2 years as a data scientist on Google’s Search Ads Metrics team. Before joining Google, Alok spent 8 years at ESPN, where he contributed significantly to the use of analytical content across all media platforms. As a founding member of ESPN’s Sports Analytics Team, he played an integral role in the development and usage of metrics such as Total QBR and various team rankings across pro and college sports, and eventually helped lead the team as it expanded and increased scope. Alok is originally from Cheshire, CT, and earned a BA/MA in statistics from Boston University in 2008. He lives in Mountain View, CA, with his wife and son.