There is often a subtle difference between why a team wins a particular match and why they might win a match in the future. Sometimes teams win games because of hugely significant and obviously important events, such as red cards for their opponents or penalty kicks awarded. However, these events tend to occur rarely and predicting when they may feature in the future cannot be done with any great confidence.
So to predict the likely outcome of a future sporting event, we need to use more common events, which correlate well with both themselves and match outcome. We are looking at goals or shots if you prefer in football and yardage stats in the NFL. In this guest post I look at how we can use logistic regression to predict game outcomes in the NFL.
Follow this link to read more.
For a similar approach for soccer, check out Zach Slaton's post at A Beautiful Numbers Game.