For those of you who are new to my method, I want to briefly describe it. The predictive standings are called that because they are a prediction of what the final standings will be at the end of the season. The method that I'm using is a derivative of the Bill James Pythagorean Record "theorem" which basically takes the runs scored and the runs against to produce a theoretical winning percentage, as opposed to the standard winning percentage (which is simply wins/total games played).

Here are the formulas used, for those who might be interested:

- Predicted Winning Percentage = Runs
^{2}/(runs^{2}+runs against^{2}) - Predicted Wins = Actual Wins + (Predicted Winning Percentage x Games Remaining)
- Predicted Losses = 162 - Predicted Wins
- Over-Performance Index = (Actual Winning Percentage - Predicted Winning Percentage) x 162

So if the Over-Performance Index (OPI) is a positive number, that means that the team is over performing their prediction of games won. If the OPI is negative, then the team is under-performing relative to their prediction. This number is actually very helpful in judging the predictive standings: if a team has a high positive OPI, then that means that they are likely to start performing more poorly in reality, once "luck" or whatever other factors this represents factor out. If a team has a high negative OPI, then that means that the record of the team should improve. If the OPI varies between -2 and +2, the team you're seeing is basically the team you're getting.

In another post, I will explain the differences between my simple version and other predictive standing methods. Virtually all of them yield nearly the same results, so you can expect that my argument for my method will encompass Occam's Razor.

## No comments:

Post a Comment