Hi wasn't sure if this would get an answer here, but it can't hurt to try.
I use lots of sports stats to calculate and predict results. Well all goes well but when I look for new things to predict, I am not really sure that I use the data my stats return in the best way. For instance lets say I run a particular theory over which I get 5000 results of a 50% event, and my data shows me a return of 54%. Making my theory correct to predict an event by 4%. I am sure at 100 samples my theory would be inconclusive. Due to the lack of data.
So my question is, in an event that is 50/50 how many samples would I need to make my theory somewhat conclusive. Like for example at 15000 samples at 54% where I need a 52% result to be conclusive. Is there a formula that would tell me if this would be enough samples to where there would be less than a 2% error. Thanks in advance Andy
Bookmarks