NFL and NCAAF spread to ML conversion

[Igetp2s]

Can someone please give a rationale for why NFL favorites are more likely to win outright than NCAA favorites when they are favored by the same spread?

I don't doubt that the stats say it is so, I was just thinking about it and couldn't really come up with any theories as to why.

Its exactly the opposite with basketball, where NBA favorites win less frequently than college favorites given the same spread.

[tomcowley]

Totals

[Igetp2s]

Can you elaborate on that? I would have thought I higher total would increase the favorites chances of winning, since there is less likely to be a fluke play causing the dog to win.

Isn't a higher total comparable to using a larger sample size in a study? Just as an extreme example, if you expect a total of 7, wouldn't you think the dog has a much better chance of winning outiright just because of luck?

[Rufus]

A higher total means that more points are expected to be scored overall. Holding point spread constant, that means that each team is expected to score more points.

For example, if Team A is -7 with a total of 35, you can infer that Team A is expected to score 21 and Team B 14. If that total was 65, those numbers become 36 and 29. In the first scenario, Team A has a larger edge in terms of the % of points being scored (21/35, or 60%), while in the second, Team B's edge is smaller (36/65, or about 55%). So while Team A is favored by the same amount in each scenario, I would say they are much more likely to win in the first scenario.

[Igetp2s]

I understand your point, but am not totally convinced.

I have been doing a lot of reading on sabermetrics recently for baseball, and it has a lot to with with eliminating the "luck" factor from the expected score.

I would have expected the fewer the scores, the increased randomness of who wins the game, while with an increase in scoring, the luck factor would even out more, and the team that is truly more skilled should be more likely to win the game.

[Data]

I understand your point, but am not totally convinced.

I have been doing a lot of reading on sabermetrics recently for baseball, and it has a lot to with with eliminating the "luck" factor from the expected score.

I would have expected the fewer the scores, the increased randomness of who wins the game, while with an increase in scoring, the luck factor would even out more, and the team that is truly more skilled should be more likely to win the game.

The luck is a less of a factor if the game is longer, not if the total is higher.

[Peep]

I think the NFL lines are tighter too.

I remember looking at totals in baseball, I don't think the affected the fav/dog ratio, at least not in relation to the line.