Standard Deviation -- how do I interpret? [Archive]

Henke

01-27-2004, 03:17 AM

Standard deviation is not really a measurement of how good you're playing, a better description would be a measurement of short term luck. However, if you're playing aggressively your SD will be higher (and the other way around). I don't know from the top of my head what values for SD that's "standard" for good players, but if you search the forums, you'll probably find something.

DeucesUp

01-29-2004, 03:07 AM

Standard deviation will give you an idea of what kind of fluctuations you can expect in your short term results. It is affected by the style you play and more so the style of the players at your table(s).

It is not, per se, related to your ability though a side effect of playing tight and playing in passive games (generally good characteristics) is a lower SD. However, playing aggressively (generally good also) will usually raise your SD. So SD is not a very useful statistic for evaluating play.

Normal online games will result in a typically standard deviation of 10-20 big bets per hour per table, with 15 being a reasonable average. The meaning of this is:

If you are break-even player in the long run, then each hour of play you would expect to be between a 15 BB loss and a 15 BB win about 2/3 of the time. About 1/6 of the hours you play would be losses of more than 15 BB and 1/6 of the hours would be wins of more than 15 BB.

The size of the standard deviations grow as the square root of the number of hours. Thus after a 4 hour session you would expect to be between a 15*sqrt(4) = 30 BB loss and a 30 BB win two-thirds of the time. With one-sixth of the 4 hour sessions bigger than 30 BB losses and one-sixth bigger than 30 BB wins.

Start experiementing with these numbers and you'll see why bad (loosing) players can win so much money over a fairly long period of time and why good players can experience such long losing streaks. It's all in the math.