[Levarkin]

I'd actually like to re-open this debate if anyone is interested. I was basing my thoughts on the WOO's analysis, and he's certainly right about the odds of favourites doing so well assuming a 50% probability. If the faves had a 50% chance of winning each of the regular season's 256 games, then there was a 1.38% chance that faves would win 144 or more times. Unlikely, but not extrememly unlikely, and not too unlikely at all after 24 seasons.

The problem is that favourites don't really have a 50% chance of winning: in the 21 seasons from 1983 to 2003, faves only cracked the 50% mark three times, and even then maxed out at 53%. The average for faves over this period was 46.92%. If we assume these are the true odds of a favourite winning, then the chances of faves winning 144 or more games drops precipitously to 0.05%. There's a 98.80% chance that a 0.05% chance won't hit over 24 trials, as the 24 seasons since 1983 (the ones for which I have records) represents.

So now I'm a little uncomfortable chalking up the performance of faves in 2005 to statistical aberration. It's still possible, just very unlikely. So a question to those who know the NFL better than I do: did anything change in 2005, either on field or within the books? Does anyone know how professionals explained the performance of favourites last year? Because good luck might just not cut it.


Caveat: I'm not a statistician, though I think between Excel and my old A-level textbook that I've got this right. I'm certainly open to the possibility of correction though.