So my PT tells me that i have an STDV of 40 PTBB/100, it's a sample of 275K hands. and it was the same result for 50K hands as well. is this standard? do you think PT calculates STDV properly?

Yes thats normal... because SD measures the dispersion in your winrates per 100 hands (or per hour, maybe even per hand)...



Where Xi is the winrate per sample of 100 hands (or per hour or per hand) and Xbar is not a nudy bar but the average over all samples.

Say you have played 300 hands... Then you've got 3 samples of 100 hands. The winrate over that 300 hands sample is the mean or X-bar and the winrate per sample of 100 hands is Xi.

So when you play more hands... you've got a lot more (Xi-Xbar)²... but you devide by a bigger number of 100 hands samples... (since you played more)... so the numerator gets bigger and the denominator gets bigger... so your SD stays pretty much the same...

Remember that this is the SD per 100 hands. Now, winrate and variance are additive... This means that the winrate over 1000 hands is simply the winrate over 100 hands times 10. For variance the same... it's the variance over 100 hands times 10.

WR/1000h = WR/100h * 10
VAR/1000h = VAR/100h * 10

And since sqrt(VAR) = SD or SD² = VAR:

SQ/1000h = SD/100h * sqrt(10)

So for a sample 10 times larger, the standard deviation increasing less then proportional, in fact: with the square root of 10 in this case.

Maybe this is what threw you a little bit off, maybe this is what got you confused. But in PT or HM we're always talking about SD per 100 hands, and that pretty much stays the same if you don't make radical changes in your play.

thanx for the clarification. yep that's what threw me off coz i was thinking if i'm playing 600 hands per hour then my SD is 6 times larger for that sample. but now looking at the formula i see how my SD per hour increases by 6/sqrt(6). still looks huge, but not as great as i imagined.