|
|
#62
09-12-2007, 02:57 AM
|
|||
|
|||
Re: 300/600 (WARNING: limit hold\'em)
[ QUOTE ]
Filtered Pokertracker for hands where 5-10 see the flop. I am at 5 BB/100 and I don't think that's particularly atypical or "running good". (I thought it would be higher because I ran higher than that at microlimits). Let's move that up to 6 BB/100 because my Pokertracker sample is at much lower limits (and thus much higher rake) than I or anyone else plays live. Let's assume 600 hands. That's 36 BB expected. Variance is something like 16BB/100, no? My Pokertracker will take forever if I try to switch tabs so I won't look. Of course this does not represent variance for 5-handed flops, but it's a starting point. We also know full ring variance is less than shorthand, so I would expect 5-handed flop to be even worse. Extrapolating this to 600 hands in my groggy midnight state, gives something close to the original 36 BB in expectation. Therefore if you break even for the session you will be about a standard dev below. Now this is somewhat likely. However, note three things: (1) Breakeven or close to breakeven is not "coming away loser". You would need to be 2 standard devs below (2) Live is going to have less variance (3) There's a tendency to not leave a good game while stuck. This is going to be a big big effect. In summary, the Central Limit Theorem is a powerful powerful thing and works quicker than you would expect. If anyone has read this far I will be amazed. [/ QUOTE ] this is all very dumb. 
|
|
#64
09-12-2007, 03:39 AM
|
|||
|
|||
|
Re: 300/600 (WARNING: limit hold\'em)
[ QUOTE ]
Variance is something like 16BB/100, no? [/ QUOTE ] I am not a highstakes limit player but this seems very wrong, and playing hands where many people see a flop should probably increase the variance.
|
|
#66
09-12-2007, 03:56 AM
|
|||
|
|||
|
Re: 300/600 (WARNING: limit hold\'em)
[ QUOTE ]
[ QUOTE ] even with replying in your own threads 20-30 times, accumulating 16k posts is absurd [/ QUOTE ] 100% agree [/ QUOTE ] tenfold the ridiculousness is accumulating 16k high-quality posts
|
|
#69
09-12-2007, 06:59 PM
|
|||
|
|||
|
Re: 300/600 (WARNING: limit hold\'em)
[ QUOTE ]
[ QUOTE ] Variance is something like 16BB/100, no? [/ QUOTE ] I am not a highstakes limit player but this seems very wrong, and playing hands where many people see a flop should probably increase the variance. [/ QUOTE ] That is a good intuition but probably not correct, because why is full-ring less variance than SH which is less variance than HU? Variance is either 16 BB/100 or 18 BB/100 (my recollection is SH is 16 and fullring is 18. I think multiway flops may lower that because of the above trend), regardless variance times square root of 6 is going to be something comparable to 36. Also, by filtering for flops 5 or more you are probably underestimating your winrate because lots of these hands will be from the blinds. This thread is making me think I am way too conservative about bankroll management and taking shots IF the game is good at a much higher limit.
|
|
#70
09-12-2007, 07:00 PM
|
|||
|
|||
|
Re: 300/600 (WARNING: limit hold\'em)
[ QUOTE ]
[ QUOTE ] Filtered Pokertracker for hands where 5-10 see the flop. I am at 5 BB/100 and I don't think that's particularly atypical or "running good". (I thought it would be higher because I ran higher than that at microlimits). Let's move that up to 6 BB/100 because my Pokertracker sample is at much lower limits (and thus much higher rake) than I or anyone else plays live. Let's assume 600 hands. That's 36 BB expected. Variance is something like 16BB/100, no? My Pokertracker will take forever if I try to switch tabs so I won't look. Of course this does not represent variance for 5-handed flops, but it's a starting point. We also know full ring variance is less than shorthand, so I would expect 5-handed flop to be even worse. Extrapolating this to 600 hands in my groggy midnight state, gives something close to the original 36 BB in expectation. Therefore if you break even for the session you will be about a standard dev below. Now this is somewhat likely. However, note three things: (1) Breakeven or close to breakeven is not "coming away loser". You would need to be 2 standard devs below (2) Live is going to have less variance (3) There's a tendency to not leave a good game while stuck. This is going to be a big big effect. In summary, the Central Limit Theorem is a powerful powerful thing and works quicker than you would expect. If anyone has read this far I will be amazed. [/ QUOTE ] this is all very dumb. [/ QUOTE ] [img]/images/graemlins/frown.gif[/img]
|
 |
|
|
Linear Mode
