What's *your* Standard Deviation?

[cnfuzzd]

ok, i thought this yesterday, but thought someone else would surely say something.


redheads. particularly redheads who looked pissed off. and in terms of BB's, it usually runs in the 20 neighborhood.


sorry. im making my drunken sweep of the forums.....

peace

john nickle

[spamuell]

How many hands do you need for this to converge?

Mine is 18BB/100 at 2/4.

I only have about 8k hands at 3/6 because I'm lazy and it's 13BB/100. Is that reliable?

[Dick in Phoenix]

8K hands is probably plenty to give you a good reading on SD. I remember reading, probably in Mason's works, that the SD of your sample converges rapidly to your "true" SD, much much faster than the mean does.

Maybe Mason could jump in here and give some guidelines for sufficient sample space for SD.

Dick

[Dick in Phoenix]

I had to convert B&M statistics ... I had 350 hours, so about 10K hands, at 4-8, with a SD of 7 BB. Converting from 30 hands/hour to 100 meant multiplying by the square root of 3, resulting in 12 BB/100.

For the nitpickers, yes, it should have been
sqr(3.33333...) but I had to round to an integer anyway.

Dick

[holdemfan]

Interesting, I never looked at this stat.
15/30= 12.98
5/10= 13.87
3/6= 13.22
2/4= 14.26
1/2= 22.86
And the exciting one!!!
NL .50= 41.35

[Octopus]

[ QUOTE ]
8K hands is probably plenty to give you a good reading on SD. I remember reading, probably in Mason's works, that the SD of your sample converges rapidly to your "true" SD, much much faster than the mean does.

[/ QUOTE ]

SD converges very quickly (compared to win rate), which is why the distribution of the answers to this post is so tight. If you were to ask about win rate, the 90th percentile would be several times the 10th percentile (assuming the 10th percentile was positive), not just 50% higher like in this question. (This would be true even if our true win rates were all identical.)

My intuition (if anyone cares) is that this number is driven first by how loose you and (especially) your opponents are, and second by how aggressive you and (especially) they are. My guess is that the SD of the most aggressive people here is not meaningfully different from the average. (Of course, I've been wrong about this sort of thing before.)

[Octopus]

[ QUOTE ]
My guess is that the SD of the most aggressive people here is not meaningfully different from the average. (Of course, I've been wrong about this sort of thing before.)

[/ QUOTE ]

SMALL SAMPLE SIZE WARNING: On the other hand, I just found a player in my database with an SD of 3.5 over 300 hands.

[sfer]

Grigsa, can you compile the results by limit? I want to see how tightly the ranges correspond to different limits. I would guess there's quite a bit of distinction between 2/4, 3/6 and 5/10, and almost none between 2+2ers at each limit.

Danka.

[MagicRat]

For what it's worth... [img]/images/graemlins/smile.gif[/img]

PP 1/2 (8878 hands): 14.7
PP 2/4 (4567 hands): 14.7 (that seems weird!)
PP 3/6 (1436 hands): 17.1 (hmmm, good or bad or normal that it's higher?? sample size WAY too small?)

[DeezNutz3]

mine is 15.5 for full tables and 22 for 6 max anyone else with some 6 max numbers??