How do you know the spread is legit?

[reno cool]

Lets say we take a year of NFL.

group A: We have the actual games with the actual spread.

group B: We take all the spreads from all the games but assign them randomly.

Based on the results of all the games, what criteria would you use to determine which group had the legit or accurate spread?

[xyz]

I would compute the variance of the actual gap between favorite and dog vs the spread for two groups separately. The one with the lower variance should be the one with the legit spread.

[Dark Horse]

You need to calculate your own line.

If I didn't have that basic info, in your scenario I would simply look at the juice and line movement.



-------------------------------------
Mathematicians live in perfect isolation; embracing that perfection, the isolation is the problem.

[reno cool]

Thanks, for the responses. I guess in the random scenario the dogs would cover a very large % of the time. Not sure anymore what I was trying to get at with this question. I guess I was wondering if there are objective reasons to assume bookmakers lines are fairly accurate?

[xyz]

Thanks, for the responses. I guess in the random scenario the dogs would cover a very large % of the time. Not sure anymore what I was trying to get at with this question. I guess I was wondering if there are objective reasons to assume bookmakers lines are fairly accurate?

I don't agree that the dogs would cover a very large % of the time in the random scenario. Assume the real spread is 7 points, it is equally likely that in the random scenario the spread becomes 10 or 4.

[reno cool]

I don't agree that the dogs would cover a very large % of the time in the random scenario. Assume the real spread is 7 points, it is equally likely that in the random scenario the spread becomes 10 or 4.

I believe the average line would be a pick. So if you bet an underdog you're more likely to win because you're getting points for nothing.

[Data]

I don't agree that the dogs would cover a very large % of the time in the random scenario. Assume the real spread is 7 points, it is equally likely that in the random scenario the spread becomes 10 or 4.

Think what happens if you replace 10 and 4 with 15 and -1.

[xyz]

Think what happens if you replace 10 and 4 with 15 and -1.

Thanks for the hint. I interpreted the proposal as randomly assigning spreads from the set of spreads in the original group. So that dogs would not become favorites after the random assignment of spread. For example, the original group has:

Fav Spread Dog
A 7 B
C 1 D
E 3 F

The random group can be
Fav Spread Dog
A 3 B
C 7 D
E 1 F