A sports gambling theory

I agree. That is true. THey overvalue traditional powers from time to time. Like Wisconsin this year was not good and Northwestern D was pretty darn good. I bet on NU to cover that +10 as well

 

That's awesome work. Thanks!

My theory isn't that the underdog is more likely to cover the point spread. It's that the underdog is a bargain on the moneyline, in such instances (high O/U) because the moneyline doesn't necessarily adjust for the larger sample size of points expected in the game.

 

Here's an example. In Week 2 of this NFL season, the Saints were favored by 9 over the Bucs, with an O/U of 47. So the predicted score was New Orleans 28, Tampa Bay 19.

According to the NFL Pythagorean formula, the Bucs had a 29 percent chance of winning the game, assuming that would be the average score between the two teams.

But the money line, at +375, was set at a 21 percent Tampa Bay win expectation rate.

So the Bucs were undervalued on the money line. You only have to win 21 percent of the time at +375 to turn a profit, and this team was expected to win 29 percent of the time this particular game would be played, according to Vegas.

Final score: Bucs 26, Saints 19

 

If there's anything we can all agree on, it's that Vegas is always vulnerable to a "system." Look at that how that place is crumbling.

Dick, while your theory makes a bit of sense on the surface, in five pages of this thread it only seems to be coming back a dud. It looks like the best you could hope for, betting $100 on each of 100 games, is to win a few hundred bucks.

 

Dick, if this was exploitable don't you think some high-money gamblers would have already figured it out? And the casinos would have adjusted?

 

I love gambling theories, and I like this concept. Maybe the hole in it is whether the NFL Pythag is a true barometer of winning expectation.

 

Through nine games of the 2015 season: Bet $900, Cash on Hand: $1,080

ROI: +$180

 

Except in most games, Pythagorean and the Moneyline align.

 

OK, now I've given that a go, too.

Suppose you bet $100 on the underdog ML everytime the O/U was at or above the 90th percentile. Over my sample, you'd have made 1,076 wagers and would have won on 301 of those. Your total winnings would have been $71,090 and your total losses would have been $76,300.

 

All right, Dick, quant's giving you the go-ahead -- you have to spend money to make (less) money (than you spent)!

 

Yep. I think they do adjust it sometimes. You have to calculate it game-by-game.

 

Let's say he's right, and the result is counter-intuitive. Well, that's exploitable, too!

Quant, what would be the result if you bet the favorite on the moneyline in those games?