standard deviation

[kennyg]

Could someone please tell me exactly what standard deviation is...what are good numbers and how I'm supposed to use this knowlegde to my advantage?? thanks!

[piersmajestyk]

Kenny,

Standard deviation is a measure of the variation in your game poker wise. With it you can compute needed bankroll requirements in conjunction with your win rate for a particular limit. If you have Mason Malmuth's Essays on Poker he gives a good analysis of the various aspects of Standard deviation and how it can be used. I don't remember all the equations needed to compute it but if you search under my name at UPF you can find a post by me that does deal with various aspects of SD, Win Rate and Bank Roll requirements.

[iceman5]

We will be posting one of the best articles ever written on this subject on the front page in the next day or so.

[kennyg]

Thanks Piers!

I always though deviation was something along those lines but wasn't sure.

[Felonius_Monk]

Kenny,

With regard to how it can help your game i'm not really sure; if you have a large bankroll for the games you play i don't think it'll ever be a problem (except with a larger SD you will experience potentially longer and nastier downswings).

Statistically speaking, standard deviation is one way to measure the average deviation from the mean, over a set of data. For instance, take this dataset: 1, 5, 9. Now, the average of this three figures is 5 but the first and third numbers deviate from the mean by 4; standard deviation is higher than in this dataset: 4,5,6. The averages are the same (5) but the SD is four times higher in the first dataset.

Now, over a period of poker play (whether it's regarded as a dataset of "winnings each hour of play", "winnings each session", "winnings per hand" etc) the average you win per 100 hands will be your BB/100, but your standard deviation measures the average amount each point in the dataset differs from that figure. For instance, if you play three hours, first hour lose 400, second lose 200, third win 800, then you're played 3 hours and won $200. Your SD will be very high. If you play 3 hours and win 100, win 20, win 80, you won $200 for an identical hourly rate of $66/hr but with a much lower standard deviation. Your bankroll curve will be smoother and you'll have less vicious downswings and also less monster win sessions.

Monk
xxxxx

[kennyg]

Very well thought out and explained Monk. Thanks for spending all that time to help me understand this concept. I appreicate it.