Making real-time predictions for NBA basketball games by combining the historical data and bookmaker’s betting line

The paper presents a gamma process based model for the total points processes of NBA basketball matches. This model obtains a useful formula for the in-play prediction. What is more, we employ the bookmaker’s betting line to adjust the original gamma process model. The out-of-sample forecasting performances are evaluated, and more profoundly, this model can produce a positive return on the over–under betting market. Besides, our model has an application in monitoring the betting market, which may be useful to bettors.