Gophers open as 8 point underdogs against Auburn.

[A_Slab_of_Bacon]

Bet generation is basically built into predictive models. There are so many games where a 4-4 team is favored over a 7-1 team and get money flowing. Or the “hot” team is an underdog to the “cold” team.

When Wisconsin was favored over us, how many dudes at a bachelor party in Vegas sprinted to the MGM sportsbook and said, “wow Minnesota is 10-1 and at home and Wisconsin looked like crap a few weeks ago against Illinois!” Wisconsin still got some action but Minnesota took 60% in spread and money line and lost both.

Yeah that makes sense.

Odds makers pushing on the scales for other purposes... that just seems like a good way to get destroyed. Especially today with so many smart folks with access to serious computing power.

I recall reading some stories about the weekly fantasy football for money type sites being dominated by a handful of folks running numbers themselves raking in the cash over the same bachelor party guys who think they know something.

 

[Some guy]

Pick spreads or winners of games?

I think the casual fan can pick Alabama to beat a bad Vanderbilt just as well as any rating. Spread is a different story.

Anyways, the point is that these spreads come from something similar to Sagarin. If you asked the casual fan what the spread would’ve been, any non-SEC homer would’ve told you it’s a pick-em maybe even Gophers favored because the Gophers are 10-2.

Pick winners of the games

you know the thing that actually matters?


Did you know that if the spread is 10 and the team wins by 5 it’s still a win.
And if the spread is 10 and the team wins by 15 it’s still a win.
in neither case did Vegas or sagarin nail the spread

 

[RickDornsby]

Pick winners of the games

you know the thing that actually matters?


Did you know that if the spread is 10 and the team wins by 5 it’s still a win.
And if the spread is 10 and the team wins by 15 it’s still a win.
in neither case did Vegas or sagarin nail the spread

Predictive models also assign probability of winning the game outright (basically correlated to spread).

If you grant that the predictive models are the best at picking an opening spread, wouldn’t you also agree that that same money line assigned giving Auburn about a 70% chance to win with the 7.5-8 point spread as accurate?

And of course the spread is missed. If the Gophers were favored by 7 points every game this year, they could win by 3 a bunch of times and 11 other times and the spread is pretty good.

Since 2003, college football teams favored by 8 points have won by an average of...8.3 points. The predictive models are pretty good.

 

[Some guy]

Predictive models also assign probability of winning the game outright (basically correlated to spread).

If you grant that the predictive models are the best at picking an opening spread, wouldn’t you also agree that that same money line assigned giving Auburn about a 70% chance to win with the 7.5-8 point spread as accurate?

And of course the spread is missed. If the Gophers were favored by 7 points every game this year, they could win by 3 a bunch of times and 11 other times and the spread is pretty good.

Since 2003, college football teams favored by 8 points have won by an average of...8.3 points. The predictive models are pretty good.

I don’t disagree that predictive models are pretty decent.

what I disagree with is that they matter.
That average voter could pick games equivalent to Auburn Minnesota correctly more than 70% of the time too

 

[Maverick]

Predictive models also assign probability of winning the game outright (basically correlated to spread).

If you grant that the predictive models are the best at picking an opening spread, wouldn’t you also agree that that same money line assigned giving Auburn about a 70% chance to win with the 7.5-8 point spread as accurate?

And of course the spread is missed. If the Gophers were favored by 7 points every game this year, they could win by 3 a bunch of times and 11 other times and the spread is pretty good.

Since 2003, college football teams favored by 8 points have won by an average of...8.3 points. The predictive models are pretty good.

yeah but... why don't they use stars??