In-game win probability models for Canadian football

ABSTRACT

This article presents in-game win probability models for Canadian football. Play-by-play and wagering data for games from the Canadian Football League for the 2015 to 2019 seasons is used to create logistic regression and gradient boosting models. Models with and without the effect of pregame spread and total (over/under) data are presented and discussed. The resulting win probability models are well-calibrated and can be used to support in-game decision-making, review coaching decisions, estimate the magnitude of team “comebacks”, and potentially identify in-game wagering opportunities. An R Shiny application is provided to allow for estimation of in-game win probability for user-provided game state inputs. Opportunities for future work are identified and described.

KEYWORDS:

• Sports
• Predictive Analytics
• Decision-making

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.