For the math wizards.......a little help please

In sports betting, using a point spread, not a money line, what would be a statistically significant sample size to make one believe that the win/loss rate over that number would/may reasonably hold up going forward? Thanks in advance for the help!

Quote: steeldco

In sports betting, using a point spread, not a money line, what would be a statistically significant sample size to make one believe that the win/loss rate over that number would/may reasonably hold up going forward? Thanks in advance for the help!

Many variables in your question need defining before even a math expert can cogently answer you. The sample size might not need to be that large if the win rate is very high. Example... if you get 90 out of 100 you would be pretty sure you have an advantage. If the advantage is small, 51 out of 100 would tell you virtually nothing. However, if you win 51,000,000 out of 100,000,000 that would tell you more. And your word 'reasonably' can have many different interpretations.

If you can give one specific example we could weigh in on it...

Luv the internet. I found a site (SportsInsights) that stated it was 2,000 games for on a 57% win rate. I'll assume that to be correct. Thanks!

Quote: SOOPOO

Many variables in your question need defining before even a math expert can cogently answer you. The sample size might not need to be that large if the win rate is very high. Example... if you get 90 out of 100 you would be pretty sure you have an advantage. If the advantage is small, 51 out of 100 would tell you virtually nothing. However, if you win 51,000,000 out of 100,000,000 that would tell you more. And your word 'reasonably' can have many different interpretations.

If you can give one specific example we could weigh in on it...

SOOPOO, thanks for responding. I looked it up using the internet (DUH!). Sometimes I'm apparently not so bright......
Anyway, the answer is in the prior post.

Thanks again though!