For those of you that read my

thread on variance, I've made a simple calculator that allows you to input your winrate, SD/100, and sample size, and will calculate whether or not that sample size is adequate.

Here's the link:

http://www.shankybotprofiles.com/samplesize.phpNote:

- If you put in a negative winrate, the calculator will switch to the number of hands needed to prove it's a loser

- SD/100 is also sample size dependent. If you have a very small sample size, then your standard deviation may be much lower if you simply haven't experienced much variance yet. In my experience, the standard deviation is fairly close to being accurate by 20-30k hands. Just be aware that you could have a situation where you have 5-10k hands with a low standard deviation and the calculator only says you need 15k. This will almost always be wrong!. Even a profile that is absolutely crushing with very low variance will need at least 20-30k hands.

For those of you who don't want to read the thread, here are the cliff notes:

Cliff Notes:

- Standard deviation is the square root of variance (which tells us how "swingy" a profile is)
- Everyone should report standard deviation for profiles along with winrate and sample size (only winrate is a more important stat for determining profile performance).
- Sample size required to declare a profile a winner is specific only to that profile and depends on the standard deviation and winrate
- There is no one sample size that covers all profiles. High winrate and low SD/100, and 50k hands is enough. Winrate of 0.5BB/100 & SD/100 of 45BB and you need 3 million hands.