Stars 2/4 6-max bankroll?

[J5983]

I thought the swings were higher because they're all crazy? And the calling stations come out in full force.

[Dov]

[ QUOTE ]
So, exactly how large do you think the typical SD is in short-handed play, and how large do you think the typical SD is in full games? Please specify figures in BB/100.

[/ QUOTE ]

I am not nearly as accomplished as you are in probability or mathematics. The only way I could get a figure for you would be from my PT database which I am not in front of now anyway.

I will give you some numbers when I get home, though. I put up mine, and the next 3 players who I have the most hands on.

[ QUOTE ]
The above passage seems to suggest you think the standard deviation should be about twice as great. I don't believe that at all.

[/ QUOTE ]

I'm not sure how much it should change the standard deviation.

It is true, however, that you should be playing almost twice as many hands in short (6 max) games, and you will be playing them much more aggressively.

Even with my limited understanding, I know this will increase the standard deviation significantly.

Let me ask you like this:

If you play 10/20 full, do you expect your SD to be higher, lower, or the same as 10/20 6 max?

If they aren't the same, then why are they different? By how much?

[Dov]

[ QUOTE ]
I thought the swings were higher because they're all crazy? And the calling stations come out in full force.

[/ QUOTE ]

Not all of them are crazy, just some of them. [img]/images/graemlins/grin.gif[/img]

Actually, the calling stations reduce your swings, if you stop trying to bluff them.

I love calling stations.

It's the LAGs who keep raising when you might have the best hand that increase your SD.

[uuDevil]

Here are some sources for SD values:

SH Poll

Micro Poll

SS Poll

[pzhon]

[ QUOTE ]
Here are some sources for SD values:

[/ QUOTE ]
I was thinking of those polls (among other things), except I have never seen the results of the short-handed poll. Now the archive server tells me I'm not logged in, and I need to log in to vote, and I can't see the results until I vote. If anyone can see the results, please copy them to this thread. Thanks.

[BluffTHIS!]

[ QUOTE ]
[ QUOTE ]
You honestly think playing more hands doesn't give you more variance?

[/ QUOTE ]
Not per 100 hands, and that is what matters for sizing a bankroll.

[/ QUOTE ]

You know pzhon, just having a knowledge of advanced statistics doesn't mean you actually are capable of logically considering and formulating a problem correctly. The whole point is that when playing any shorter handed game you are playing way more hands per 100 than in a full game which necessarily has to increase your variance regardless of your edge. And if you think you can just play weak-tight and NOT play more hands in a short-handed game then you will lose.

[pzhon]

[ QUOTE ]

It is true, however, that you should be playing almost twice as many hands in short (6 max) games, and you will be playing them much more aggressively.

[/ QUOTE ]
It is obvious that you play more hands short-handed. I had that in mind from my first post on this thread. It will save time if you don't assume I am missing the obvious.

Your variance is greater in multi-way pots, which are more frequent when there are more players. You can pump big draws for value and you get great pot-odds on weak draws. Does the fact that you play more hands dominate the increase in multiway pots? The numbers should tell us.

[ QUOTE ]

Even with my limited understanding, I know this will increase the standard deviation significantly.

[/ QUOTE ]
Your intuition may be wrong by a lot. That is why I am asking for numbers.

In the HUSH forum, many people have stated that there is a minimal difference between the SD of short-handed tables and full tables. I trust PokerTracker data far more than I trust your intuition. From what people have posted in the HUSH forum, I think the SD might be about 10-15% higher, (15-20 BB/100 instead of 13-18 BB/100), which would mean that if you have the same win rate, you need a bankroll that is about 20-30% larger. The difference in win rates can easily be much larger than this.

I believe the huge downswings people discuss primarily come from lowered win rates, not increased volatility. People also underestimate how large the swings can be in full games.

[Dov]

OK, I'm home now.

I'm only going to post my stats though, because I need to get some hands in tonight and I got home late.

5/10 6 max

Hands 48167
VPIP 29.77
PFR 18.14
BB/100 1.89
SD 22.45 BB/100

10/20 6 max
Hands 729 (I know - small sample size) (I just took a shot and got killed)
VPIP 27.98
PFR 16.74
BB/100 ( -4.88)
SD 13.35 BB/100

Sorry about the small sample size for 10/20. I couldn't take anymore beating that day. [img]/images/graemlins/crazy.gif[/img]

[pzhon]

[ QUOTE ]
You know pzhon, just having a knowledge of advanced statistics doesn't mean you actually are capable of logically considering and formulating a problem correctly.

[/ QUOTE ]
That I am a mathematician does mean I am good at logical analysis and formulating problems correctly.

I understand you are still sore that I pointed out your logical errors on thread after thread, e.g., when you called me a statistics donkey and posted your bankroll formula which depends on session length but not the toughness of the game! However frustrated and embarrassed you are does not mean I am wrong here. Is your game selection this bad in poker?

[ QUOTE ]
The whole point is that when playing any shorter handed game you are playing way more hands per 100 than in a full game which necessarily has to increase your variance

[/ QUOTE ]
Playing in multiway pots increases your variance. There are more multiway pots in full games. Playing larger pots increases your variance. Pots are larger in full games. I just sampled the Party $30-$60 games, and the full tables had an average pot size of $342 = 5.7 BB, as reported by Party. The tables with 4-6 players had an average pot size of $255 = 4.3 BB. In the $5-10 games, the full tables had an average pot size of 5.4 BB, while the 6-max tables had an average pot size of 4.8 BB.

This doesn't mean I think the variance is higher in full games. There are factors pointing in opposite directions, and actual data from PokerTracker would be better than repeating the qualitative arguments. In the HUSH forum, people usually report standard deviations between 15 and 20 BB/100, which doesn't seem much greater than in full games.

[Dov]

This response is interesting to me. Intuitively, it seems like there are things missing from your analysis, but I could be wrong.

[ QUOTE ]
Playing in multiway pots increases your variance. There are more multiway pots in full games.

[/ QUOTE ]

This is not necessarily true. Many full games have only 3-4 players / pot, while many short games have 3-4 players / pot. It's just that in the short game, it's the same players over and over.

A good table in a full game (from a VPIP perspective) would be anything over 25%. A good table in a short game is anything over 30%.

25% of 10 is 2.5. 30% of 6 is 1.8.

According to this, you should have more variance in the full game. [img]/images/graemlins/crazy.gif[/img]

I don't get it.

[ QUOTE ]
Playing larger pots increases your variance. Pots are larger in full games. I just sampled the Party $30-$60 games, and the full tables had an average pot size of $342 = 5.7 BB, as reported by Party. The tables with 4-6 players had an average pot size of $255 = 4.3 BB. In the $5-10 games, the full tables had an average pot size of 5.4 BB, while the 6-max tables had an average pot size of 4.8 BB.

[/ QUOTE ]

I think the difference is that you have proportionally larger pots in the SH games. For example, the 5/10 full was 5.4 BB / 10 players = .54 BB/player on average. The 6 max was 4.8 BB / 6 players = .8 BB/player.

This is a 32.5% increase in player involvement in 6 max versus full.

Am I making a mistake here?