I'm trying to find out the confidence interval for my sample of win/loss picks. I've read a few books that mention using binomdist function and how to calculate excess wins and standard error, but I'm still not where I want to be with a definitive number. Right now I have this data:
Std Error: 42.5
Excess Wins: 43.3
I've read that to have 95% confidence you need 2 standard errors, which would be 85 excess wins so I don't have that. But I do barely have the 68% confidence you get from 1 standard error. Correct right?
And what about using binomdist function? I guess I just don't understand what the resulting % means to me. Binomdist(#wins, #samples, X, TRUE) yields a % result. I'm lost on what to enter for "X" and what that then means in the resulting % result. I read that if you enter the breakeven money win% for "X", say 52.4% for -110 vig lines, you get a resulting % of how often a random sample hits less than the breakeven 52.4%. But I'm still lost as to what that really means. Am I looking for a result that says 99.9%, ie does a higher resulting % mean more confidence? Is it the actual confidence interval? For my data, using the breakeven excess wins winrate yields a binomdist resulting % of 99.9%, seemingly proving I can be super confidence that I'm better than breakeven. But again, above we show I only have 68% based on the excess wins figures.
Help is appreciated.
Std Error: 42.5
Excess Wins: 43.3
I've read that to have 95% confidence you need 2 standard errors, which would be 85 excess wins so I don't have that. But I do barely have the 68% confidence you get from 1 standard error. Correct right?
And what about using binomdist function? I guess I just don't understand what the resulting % means to me. Binomdist(#wins, #samples, X, TRUE) yields a % result. I'm lost on what to enter for "X" and what that then means in the resulting % result. I read that if you enter the breakeven money win% for "X", say 52.4% for -110 vig lines, you get a resulting % of how often a random sample hits less than the breakeven 52.4%. But I'm still lost as to what that really means. Am I looking for a result that says 99.9%, ie does a higher resulting % mean more confidence? Is it the actual confidence interval? For my data, using the breakeven excess wins winrate yields a binomdist resulting % of 99.9%, seemingly proving I can be super confidence that I'm better than breakeven. But again, above we show I only have 68% based on the excess wins figures.
Help is appreciated.