JohnStOnge
November 25th, 2007, 06:38 PM
This is the kind of thing I've been talking to Boogs about. See how his system compares in terms of prediction to more established systems.
I looked at the 9 FCS games played this past weekend. 8 playoff games plus the Bayou Classic. Higher ranked teams according to Boogs won 4 times. Higher ranked teams according to Sagarin won 7 times (alas, the only Sagarin higher rated playoff team of a pair to lose was McNeese).
Both systems missed on the Bayou Classic and the EWU/McNeese playoff game. However, in all three cases in which the two systems disagreed (DSU/Delaware, JMU/App, Montana/Wofford), the Sagarin favorite prevailed.
That's not quite a "significant" difference by convention. The probability that Sagarin would "beat" Boogs system by chance in all three disagreements is 0.125; or one chance in 8.
But that's the idea in terms of how I think he ought to evaluate his system to validate it.
Can't rag him too much on Grambling losing to Southern because Sagarin picked Grambling for that one too.
I looked at the 9 FCS games played this past weekend. 8 playoff games plus the Bayou Classic. Higher ranked teams according to Boogs won 4 times. Higher ranked teams according to Sagarin won 7 times (alas, the only Sagarin higher rated playoff team of a pair to lose was McNeese).
Both systems missed on the Bayou Classic and the EWU/McNeese playoff game. However, in all three cases in which the two systems disagreed (DSU/Delaware, JMU/App, Montana/Wofford), the Sagarin favorite prevailed.
That's not quite a "significant" difference by convention. The probability that Sagarin would "beat" Boogs system by chance in all three disagreements is 0.125; or one chance in 8.
But that's the idea in terms of how I think he ought to evaluate his system to validate it.
Can't rag him too much on Grambling losing to Southern because Sagarin picked Grambling for that one too.