Two Plus Two Older Archives (http://archives2.twoplustwo.com/index.php)
-   Probability (http://archives2.twoplustwo.com/forumdisplay.php?f=23)
-   -   Another SD question... simple though (http://archives2.twoplustwo.com/showthread.php?t=55417)

 ZManODS 12-01-2003 11:25 PM

Another SD question... simple though

Say i calculated my Standard Deviation to be 30.7 for the last 8 sessions i played in 2/4. Yes i know this is a EXTREMELY small sample but what can i conclude from this??

Does this mean i make +-\$30.7/hr?

Also, what is a good standard deviation? (Smaller better i presume) but how small is good? I know these are silly general questions, but this sort of mathematics is interesting to me.

 SevenStuda 12-08-2003 01:12 PM

Re: Another SD question... simple though

Good standard deviation is 0.

 Analyst 12-08-2003 03:17 PM

Re: Another SD question... simple though

[ QUOTE ]
Say i calculated my Standard Deviation to be 30.7 for the last 8 sessions i played in 2/4. Yes i know this is a EXTREMELY small sample but what can i conclude from this??

Does this mean i make +-\$30.7/hr?

[/ QUOTE ]

I suspect that you can conclude very little at this point. As to what SD means, one way to think about it is the width of the "bell curve" that describes your probabalistic results. The SD needs to be on some type of normalized basis, say 30.7/hour or 30.7/100 hands, and you need to say if it's measured in \$ or BB. Assuming that an SD of 30.7 is correct (and you're right that the sample size is too small at this point, but . . . ), then you can expect that your results will fall within +/- 30.7 of the mean ~68% of the time and within +/-61.4 of the time roughly 95% of the time, assuming that you're playing for one normalized unit (that is, 1 hour or 100 hands as described above).

More generally, your results for an n-hour session will be have a mean of n*(your average/hour) with a SD of square root(n)*(your SD/hour).

 All times are GMT -4. The time now is 04:04 PM.