Will the Patriots Cover the Spread?

This week, a writer I have mildly warm feelings for profiled Kenny White, the lead oddsmaker and owner of Las Vegas Sports Consultants, the firm with the most influence over the initial Super Bowl point spread.

The spread LVSC sent out to clients (the vast majority of the sports books in Vegas) was 14 points in favor of the New England Patriots but has since slipped to 12 points. The reasons? The fact that NE QB Tom Brady hurt his ankle might have something to do with it, but it's more likely that pro bettors were looking at the Pats past ability to cover the spread. The team might have gone 18-0 in the regular season, but their record against the spread was only 10-8 (this is after going 8-0 to start the season). Still, look for the line to creep back up towards 14 as recreational bettors -- who typically bet the favorite -- start to bet en masse closer to the game.

Looking elswhere, bettors at the TradeSports prediction market are giving the Patriots a 49.5 percent chance of covering a 12.5 point spread (and an 82 percent chance of winning outright).

Another interesting source, Steve Sapra at Analytic Investors, uses what the firm calls N.F.L. Alphas -- a measure of the profitibality in betting on a team for the entire season -- and recommends picking the team with the lower alpha. Looking at the 1988-2006 N.F.L. seasons:

Our data shows some evidence that bettors systematically overestimated the likelihood of victory for the higher-alpha teams in the next season. In other words, teams that did surprisingly well in one season tended to do surprisingly poorly the next season, and vice versa...Assuming that the playoffs constitute the next season,we believe that the best strategy for making NFL postseason predictions is to favor the lower-alpha team in each respective game.

I think that assumption leaves a lot to be desired (a team's roster is nearly identical between regular season and the playoffs, you can't say the same about inter-season rosters), but since the Giants' alpha is lower than the Patriots' alpha, Sapra's method suggests the Giants are a good bet.

Barring any major injury to the Patriots between now and Sunday, I'm more inclined to join the "dumb money" and predict the Patriots will cover a 12-14 point spread for three historical reasons:

1) As White told me in the piece, the Pats do a good job covering in pleasant climates, which they're guaranteed of in a domed stadium in Arizona.
2) The last time the Patriots played in a Super Bowl with such a big spread they prevailed (although they played the part of underdogs).
3) In Super Bowl's past, heavy favorites (10 points or more) have covered 8 out of 13 times.

UPDATE
The perils of making public predictions.