In-play forecasting of win probability in One-Day International cricket: A dynamic logistic regression model

The paper presents a model for forecasting the outcomes of One-Day International cricket matches whilst the game is in progress. Our 'in-play' model is dynamic, in the sense that the parameters of the underlying logistic regression model are allowed to evolve smoothly as the match progresses. The use of this dynamic logistic regression approach reduces the number of parameters required dramatically, produces stable and intuitive forecast probabilities, and has a minimal effect on the explanatory power. Cross-validation techniques are used to identify the variables to be included in the model. We demonstrate the use of our model using two matches as examples, and compare the match result probabilities generated using our model with those from the betting market. The forecasts are similar quantitatively, a result that we take to be evidence that our modelling approach is appropriate.