Line movement math

Assume you can accurately beat the spread on 60% of your NFL bets. Meaning, you guess the correct side of where the line will move and not actually winning 60%. Example: I bet Oakland at +5 and then they move to +3. And of course we know blindly betting achieves a 50% win rate. What is the new win rate???

Unbelievably difficult question to even take a stab at because not every line movement is created equal.

Quote: SM777

Unbelievably difficult question to even take a stab at because not every line movement is created equal.

Yea. You could even lose! You correctly pick Oakland to go from 5 to 5.5 60% of the time. 40% of the time you bet them at +3 and they close at +7.5.

Basically, I think the question involves the equivalence of 1 or 2 points in the line to winning pct?

I don't know, except that it is definitely not linear. The difference between +2.5 and +3.5 is much more significant than the difference between +4.5 and +5.5, due to the typical margins by which teams beat each other.

Quote: Chodempole

And of course we know blindly betting achieves a 50% win rate.

In reality of course, with a straight ten percent line, the "House Advantage" is five percent. The huge myth played out by bookies who give credit is "you only pay when you lose". The mathematical fact is they should win five percent of the handle. So your question really revolves around overcoming a HA of five percent using line movements as a method.

Quote: Chodempole

What is the new win rate???

Less than 51%.

If someone bets every NFL opener and beats the closing spread 60% of the time, that means they are getting worse than the closing spread close to 40% of the time. That isn't going to beat the market. Just as betting blindly will win 50% of the time, betting blindly will beat any line movement 50% of the time as well. Of that one time in 10 you do get a better line than a random bet, every extra half point will only change the result about once every 50 games.

Maybe a better way to express the question is, if you beat 1000 lines by a half point, what would your win rate be? As TomG said, blindly betting achieves 50% win rate so it has to be better. What about beating it by a point, or a point and a half?

Quote: Chodempole

Maybe a better way to express the question is, if you beat 1000 lines by a half point, what would your win rate be? As TomG said, blindly betting achieves 50% win rate so it has to be better. What about beating it by a point, or a point and a half?

It depends. Half point off a three is different than half point off a five. There’s data out there on this.

Quote: Chodempole

What about beating it by a point, or a point and a half?

If someone were to bet every NFL regular season game randomly over the next four years (1024 games) and get one point better than the closing number we should expect about 15 losses to turn into wins based on the data I've seen

Quote: unJon

It depends. Half point off a three is different than half point off a five. There’s data out there on this.

Exactly. My site has data on the probability of winning getting or laying 0.5 to 7 points, given any common point spread. With a bit of work I could calculate a blended average, but what would be the value of such a number?

If someone is claiming to beat the spread by x points y% of the time, I would be skeptical. He is probably betting only games where the extra point is unlikely to make a difference, like off a point spread of 5.

Quote: Wizard

Exactly. My site has data on the probability of winning getting or laying 0.5 to 7 points, given any common point spread. With a bit of work I could calculate a blended average, but what would be the value of such a number?

If someone is claiming to beat the spread by x points y% of the time, I would be skeptical. He is probably betting only games where the extra point is unlikely to make a difference, like off a point spread of 5.

I don’t see any value in that blended number. I think these are questions the OP should be asking:

1) For each given spread, what’s the probability that the final score is within X points of the spread number. (It will vary depending on the point and looking at the historical distributions for each point is an interesting exercise).

2) What is more predictive the opening line or the closing line? (Or is there no statistically significant difference.) In other words, the distributions in #1 could be X points from opening line or X points from closing line. There’s an old sports gambler’s adage that if the point moves by 2 pts or more, bet against the crowd . . .