Quote:
Originally Posted by Garick
Shai, most of the SDev's discussed ITT have been per hour, vice per 100 hands. 34.93BB/hr is very low. GG's is 59BB/hr. I've actually never seen a long-term winner within SDev below 50BBs/hr before. Average ITT is about 100BB/hr.
Most variance calculation tools and bankroll management apps use BB/100 for some reason or another (everything I've seen anyway), and 100BB/100 is what I always hear is "normal", which converts to roughly 55BB/HR. You are right though about 34.93BB/HR being very low and unusual for a winning player.
I had mistakenly read
this post from MikeStarr where he cited a SD of 61BB/
HR where I got the units wrong and thought he had said 61BB/100, since those are the units I always use in calculations. I had been wondering how the hell he managed to win that much with 61BB/100, lol.
I have 145BB/100 @ 11.5BB/hr which converts to...
sigma_hr = 145/sqrt(100/30) = 79.4BB/hr
Guess I'm nittier than I thought.
I think you may be mistaken about the average standard deviation of winning players being 100BB/HR. If do a cursory search about standard deviations in poker you'll see 100BB/100 or even less cited over and over. For instance I just googled "normal standard deviation site:twoplustwo.com" and the first result is
this thread from 2012 and the second result
this thread from 2013 where everyone is citing 70 - 100 BB/100. Either people have gotten way LAGgier in the intervening years or people are mistakenly using BB/100 and BB/HR interchangeably. I don't think 100BB/HR is normal. Depending on hands dealt per hour that is about 180 - 200 BB / 100 (33 to 25). Which would mean "normal" is much LAGgier than I play and I'm pretty sure I am a relative maniac compared to most on this site.
I could certainly be wrong.
@GG - Standard deviation scales with the square root of sample size, not linearly, which is why you have seemingly weird standard deviations for different sample sizes (per hand, per hour, per 100, per session). If your session SD is 459.43/HR then you would average x hour sessions where 459.43 = 160.46*sqrt(x), so x = (459.43/160.46)^2 = 8.20 hours per session. Does 8.2 hour sessions sound about right?
Oh you gave the hours, you average 7.8 hours per session. Sounds close enough, your accounting of the hours or winnings may be slightly off, or...maybe the way the program converts SD to an hourly rate. Because there is no precise way to convert SD to an hourly rate since hands are dealt at different speed. Which I'm guessing is why the BB/100 unit is standard, not BB/HR. Your BB/session SD should be 100% accurate as there is no hand/hr conversion needed for that.
Last edited by Shai Hulud; 11-27-2018 at 01:13 PM.