large enough sample size to call a player or a trend successful?

[mofome]

two different questions, but what are the answers in your opinion?

[BigOrangeTitans]

minimum of 250 for me

[accuscoresucks]

variance and diffrent sport schedule i sat 750 min.

[Dark Horse]

Use Z-factor. When above 3 the angle is solid. Above 2.5 is good as well.

Z factor = wins minus losses, divided by the square root of the sample size.

[Dark Horse]

The problem with percentages is that you will not notice a trend until it jumps out. The trend gets a head start, so to speak. It may simply be a short term fluctuation. Important to only start the count from the time you notice the trend (not going back to the time that you think it started).

The problem with an arbitrary sample size of '750' is that in most cases you will be years into the project before you can start betting. Not to mention that you will hardly have any angles.

Z-factor or Z-score is the way to go. It is more reliable than percentage, because it factors in sample size.

[armyoflovers]

i was told 30 in school, so thats it for me

[thebigguy]

I think 100.