FanPost

Predictive mechanisms

Not long ago, I was involved in a debate here regarding the usefulness of SRS to predict performance. I decided that SRS (which stands for Simple Rating System) should be tested against the simplest possible predictive methods: W/L and home field advantage. So, for the remainder of the season (from the All Star break forward) I will be applying these three methods to each series in an attempt to determine if SRS is any better than either of the other methods. If people are interested, every couple of weeks I will post an update to how it is going.

So far, 30 three game series have been played, and 15 four game series.

In 3 game series so far:

SRS is 18-12
The team with the better record is 14-16
The home team is 18-12

In 4 game series:

SRS is 6-5-3
Team with better record is 7-4-4
Home team is 5-6-4

(One note about four game series: if SRS predicts a tied series, this counts as a win for SRS, however, if one team or the other wins the series, the series is thrown out for the purpose of SRS, as I felt that was more fair.)

Sample size is still quite small, but my hypothesis (that there will wind up being no appreciable difference between the three) seems to already be borne out somewhat.