|
|
#7
05-20-2007, 12:20 PM
|
|||
|
|||
Re: HU SNGs
variance is easy to quantify for HU matches as each match is a binomial(1,p) random variable. Its a known fact that you only need like 200-250 HU matches to have a VERY good estimate of your win rate. For HU SNGs you need like >1000 matches. This should answer your question about variance. If we need 4-8x more games to know about win rates in SNGs then they must have much higher variance.
|
|
|
Threaded Mode