Expected Runs in Finnish Baseball

This work proposes a discrete-time Markov chain model for the game of Finnish baseball (pesäpallo). In the model, the state of the game at the beginning of a hitter’s plate appearance is represented by the locations of the base runners as well as the number of outs. The dynamics of the game are described using discrete state transitions whose probabilities are estimated from play-by-play data. The runs are counted using special pass-through states which are visited whenever runs are scored. The model can be used to, e.g., estimate the expected number of runs for the rest of the inning, the added value produced by the plate appearances of a given player, or the team’s winning probabilities given the score and the system state.

Co-authored with Aku Kettunen.