Sport score is an app that allows users to track their favorite sports teams, leagues and competitions. The app also offers chat and messaging capabilities, as well as the option to stream goal clips. It is available on iOS and Android devices.

In some sports, winning depends on achieving a set number of points sooner than one’s opponent. In others, such as boxing and MMA, victory is achieved by knocking out or submission the opponent. In racquet sports, such as tennis and squash, points are won when the opponent fails to return the ball or shuttlecock (as in doubles).

We use a large data set of scoring events across several seasons of college and professional football, hockey and basketball to identify patterns and principles that cut across different sporting contexts. We find that the tempo of scoring events, as measured by the inter-arrival times of adjacent events, closely follows a Poisson process with a sport-specific rate. Similarly, the balance of events won by each team closely follows a Bernoulli distribution with a parameter that varies with the size of the lead.

We further show that the observed tempo and balance of scoring events predict the evolution of lead-sizes with remarkable accuracy. This knowledge, combined with the information on team skill incorporated in our model, allows us to make predictions of game outcomes, even when only the first few scoring events have occurred. We further show that these predictions are more accurate than those produced by several commercial odds-makers.