Quote:
Originally Posted by basser
What has more variance, limit holdem or no limit?
Variance is a statistical measure of the degree to which data varies from average. At the same blind level, (comparing 200nl to $2/$4 limit -- both with a $2 big blind -- for example),
no-limit will have much greater variance statistically because pots can be far larger. In one session, you might be up or down 300 big blinds in no limit, but +/-300BB would be unusual, to say the least, for a limit session.
People will tell you the variance in limit is greater, but they're not really speaking mathematically. What they mean is something like "losing streaks are more frequent, longer, and less avoidable in limit." One reason is that winrates of good players are lower in limit. If you are winning 3BB/100, it is much easier to end a session for a loss than if your winrate is 15ptbb/100. The +15ptbb/100 no-limit player might play a session for +5ptbb/100 and not consider that a "losing" session, but that actually represents a bigger negative departure from the mean than a +3BB/100 limit player who has a -3BB/100 session.
Also, most people won't compare limit to no-limit at the same blinds. It will vary wildly with individual skills, but a $1/$2 no-limit player might be playing $5/$10 when he plays limit. Higher blinds makes for bigger pots and therefore higher variance, by simple arithmetic.
Attempting to make an apples-to-apples comparison is difficult, but I did an cheap overlay of two players who achieved the same $/hand winrate, one playing limit, the other no-limit. Here's the graph:
red line is a no-limit player playing 25NL 6-max for 17,000 hands
green line is a limit player mixing up .25/.50, .50/1 limit 6-max and a little .50/1 limit HU, also 17,000 hands
blue line is the mean winrate they share over the 17k hands. You can see that the red line varies much more widely around the mean.
Last edited by gedanken; 08-08-2009 at 02:06 PM.