Sport Score Models for Four Major Sports


sport score

Sport score is an app that aims to give you the latest news, scores, stats and more from NFL Football, NBA Basketball, MLB Baseball, NHL Hockey, NCAA Basketball, WNBA Basketball, EPL Soccer, La Liga Soccer and all other major leagues and competitions. The app also lets you personalize your feed with the teams and stories you care about.

Across the four studied sports, scoring tempo — when scoring events occur — is remarkably well-described by a simple Poisson process in which each second on the game clock produces a scoring event with a random probability that remains fairly constant across gameplay. In contrast, scoring balance — how often each team wins a scoring event — shows a more complex pattern. It reveals a Bernoulli process in which the probability of winning a scoring event while leading effectively increases with lead size.

The generative models developed in this study are designed to capture these patterns by modeling scoring dynamics as a first-order Markov chain, wherein past gameplay events have only a finite effect on the future outcome of an iid random variable that is generated from data. These models are computationally efficient and exhibit good predictive performance, comparable to that of commercial odds-makers.

Results from the BESS indicate that total performance differed between sports (F6,347 = 3.147, P .001) and had medium effect sizes (ep2 = 0.05). Post hoc Hochberg GT 2 comparisons revealed that the men’s soccer and baseball/softball teams performed better than the women’s soccer and basketball teams.