sutz":kxesk2h8 said:
Guys, guys, guys.....and gals.

Remember, it's a
betting line, not a prediction. It's not that Vegas thinks we'll win by ten, it's that they think nobody will bet the Cards unless they parlay a ton of points. The fact that the line hasn't moved much indicates that the American sports betting public agrees that the Cards don't have much of a chance.
You are completely right, but the initial line is still a very strong input for prediction models for perhaps non-intuitive reasons.
i) Assume that Vegas isn't trying to predict anything but is instead trying to split the money. Specifically, they are trying to split the total quantity of money rather than the total quantity of bets. Vegas is perfectly happy if 70% of bettors pick one team and 30% pick the other, so long as the money involved is 50%/50%.
ii) Assume that the amount bet by each person is highly correlated with their level of information - few can afford to bet very much if they bet poorly. Furthermore, the more confidence that people have in their side (the further the line is from their opinion), the more they will bet.
1) Ergo, If Vegas's line is "wrong", say to accommodate a lot of grandmas betting on the Cowboys, then they are at risk of having the large/smart money end up on the same side and dramatically tilting the overall balance of money. They only get a safe guaranteed profit if the money matches and if there was a long-term trend exposed in their lines (overrating/underrating team X) then it would be picked up on and Vegas might lose money.
2) Ergo, Vegas pays a lot of money to the "best" odds makers to predict the match-up close enough that they do not risk losing money. If the money starts coming in on one side too much anyway, or if things change, then they promptly move it to balance things and correct any oversights made by the initial odds-maker.
For example, the Dallas Cowboys covered the spread 49.9% of the time between 1978 and 2010. The status of being America's team with a bunch of casual bettors favoring them every week made no impact on their average spread, which instead nearly perfectly fit their actual capabilities.