Smart People

[BobboFitos]

I made a post in MHNL (where I normally post) and realized maybe probability forum was better suited for it. (Then again, maybe not, as this is not about probability [img]/images/graemlins/crazy.gif[/img]) I dont know if anyone wants to tackle the problem, as I dont really have much mathematical framework to work with, but: Many people always try to determine successful winrates. No one really knows what certain sustainable rates are, esp. for higher stakes games.

My conjecture, and it wasnt even stated that well, is one way to determine the "tops" winrate was to find the worst winrate, (over the long haul) and from that draw a conclusion that because poker is a zero sum (or negative sum, with the rake, so +Winrate would be < (-) -Lossrate) that the worst rate would match up with the best way.

I had an idea for finding that lowest rate, but once again, not really sure (it's stated in my post) if I was / am on the right path.

So, if anyone wants a hack at it, here's the Link to the original post

[ThinkQuick]

From your other post:
[ QUOTE ]
So, i'm projecting that since -7.5ptbb/100 is the worst sustainable rate, the positive of that should be the best sustainable rate.

[/ QUOTE ]

This is not obvious to me, and I don't intuitively believe it to be true. The fact that the statement has not yet been attacked to badly may be some support for it, but I think I'd need some further evidence.

[pzhon]

[ QUOTE ]
No one really knows what certain sustainable rates are, esp. for higher stakes games.

[/ QUOTE ]
I disagree. There will always be some uncertainty, but some people have played hundreds of thousands of hands. That's enough to drive the standard error from results alone down to about 1 PTBB/100. Others can pool their data, or datamine.

[ QUOTE ]
My conjecture, and it wasnt even stated that well, is one way to determine the "tops" winrate was to find the worst winrate, (over the long haul) and from that draw a conclusion that because poker is a zero sum (or negative sum, with the rake, so +Winrate would be < (-) -Lossrate) that the worst rate would match up with the best way.

[/ QUOTE ]
That doesn't work.

[img]/images/graemlins/diamond.gif[/img] There are several players at the table who may be losing, not just one. There may be several winning players, not just one. In high stakes limit games, there are often several small winners feeding off a couple of big losers.

[img]/images/graemlins/diamond.gif[/img] You haven't said anything that distinguishes high-stakes games from low-stakes games where I presume you do accept that people know their win rates, and people have win rates much higher than 7.5 PTBB/100.

[img]/images/graemlins/diamond.gif[/img] You assumed that the worst possible win rate is to lose the blinds. In fact, many players lose much more than that. Some players defend their blinds far too much, or smooth-call raises with trash, or limp with high pairs and stack off against trips, and they dump far more than 7.5 PTBB/100.

[AaronBrown]

I agree with phzon (not just on this post, in general). However, it is worth considering the relation between positive and negative win rates. Certain rates of losing will drive players from the game. You might conjecture that there will always be replacements for them, but that seems like wishful thinking.

If you could arrange your games like Frank Wallace (author of Advanced Concepts of Poker) where you are the only big winner and the rates of losers are restrained to keep the game going, you could win many times what the worst player loses. But it's pretty hard to capture all the losses at a Poker table, and it's hard to keep the loss rates just at the point where people continue losing.

I think for most situations, it's hard to win consistenly more than the worst player who continues playing long-term loses. However good you are, you will have trouble keeping a game going in which you get all the money, and you will have trouble keeping other good players from poaching on your turf.

hello AaronBrown

I really look out for your posts, as often they are an oasis of facts and rationality, and are always worth reading. Thanks.

I may take this slightly off topic, but I can't resist as it is something I have pondered upon a lot for quite some time.

Going on from what you say, could it not be that the rake (which is a casino manipulated variable) is set just so that it maximises the house's return, so that it is set to the maximum EV- for the players which will still allow growth in the business.

[BobboFitos]

Sorry it's taken me a day+ to get back to people. I was away.

[ QUOTE ]
I disagree. There will always be some uncertainty, but some people have played hundreds of thousands of hands. That's enough to drive the standard error from results alone down to about 1 PTBB/100. Others can pool their data, or datamine.

[/ QUOTE ]

I'm talking no limit win rates, not limit. The link is to the mid high no limit forum.


re: your other conjectures, it's tough for me to argue because you assumed I ment a limit game.

Mainly, I say -7.5ptbb/100 is the tops a player can sustainably lose, meaning, someone trying to play well, therefore, doesn't just drop 1000bbs (for example) then leaves, i'm saying thye play for 100 years. This statement is awkward, maybe, but it means then really it should be the worse they are, the limit approaches that number, but never makes it that far.

[pzhon]

[ QUOTE ]

it's tough for me to argue because you assumed I ment a limit game.

[/ QUOTE ]
I assumed you were talking about NL. Everything I said was based on NL. Why do you think I was talking about limit?

[ QUOTE ]
Mainly, I say -7.5ptbb/100 is the tops a player can sustainably lose, meaning, someone trying to play well, therefore, doesn't just drop 1000bbs (for example) then leaves, i'm saying thye play for 100 years.

[/ QUOTE ]
I disagree. I think some people would continue to lose much more than they would by posting and folding.

The maximum sustainable loss rate by someone who is trying to win should depend on the stack sizes. I think that with 50 BB stacks, everyone wins with AA, and perhaps KK. With 500 BB stacks, I think it is easy for some people to lose a lot with AA, but they won't start folding AA preflop any more than they fold AT with 50 BB stacks.

[BobboFitos]

[ QUOTE ]
[ QUOTE ]

it's tough for me to argue because you assumed I ment a limit game.

[/ QUOTE ]
I assumed you were talking about NL. Everything I said was based on NL. Why do you think I was talking about limit?

[/ QUOTE ]

Because 1ptbb/100 is not a good winrate. Sure, it's positive, but relatively low.
[ QUOTE ]


[ QUOTE ]
Mainly, I say -7.5ptbb/100 is the tops a player can sustainably lose, meaning, someone trying to play well, therefore, doesn't just drop 1000bbs (for example) then leaves, i'm saying thye play for 100 years.

[/ QUOTE ]
I disagree. I think some people would continue to lose much more than they would by posting and folding.

The maximum sustainable loss rate by someone who is trying to win should depend on the stack sizes. I think that with 50 BB stacks, everyone wins with AA, and perhaps KK. With 500 BB stacks, I think it is easy for some people to lose a lot with AA, but they won't start folding AA preflop any more than they fold AT with 50 BB stacks.

[/ QUOTE ]

yea. well... I agree. any idea though between correlation between max win and max loss?

[pzhon]

[ QUOTE ]
[ QUOTE ]

I assumed you were talking about NL. Everything I said was based on NL. Why do you think I was talking about limit?

[/ QUOTE ]
Because 1ptbb/100 is not a good winrate. Sure, it's positive, but relatively low.


[/ QUOTE ]
I never said anything about a 1 BB/100 win rate. "There will always be some uncertainty, but some people have played hundreds of thousands of hands. That's enough to drive the standard error from results alone down to about 1 PTBB/100. Others can pool their data, or datamine."

"Standard error" and "standard deviation" refer to the variation of a quantity about its expected value, not the expected value itself.

[AaronBrown]

Thanks for the kind words.

Casinos definitely manipulate the edge in house games. Players determine how much they are willing to lose, the casino's choice is whether to take that slowly (with low edge) or quickly (with high edge). Slowly gives more value for money and wins more repeat business, but it costs money in that you need more tables and overhead.

Of course, there is a range for different types of players with different preferences. Some players like the low edge per bet but high edge per hour of craps; others like the higher edge per bet but lower edge per hour of roulette. Also, the economics are different fora monopoly casino (like Foxwoods with nothing comparable in easy driving range) versus one in a competitive strip (like Vegas or Atlantic City). Midweek is different from Saturday night.

With Poker it's simpler since the house takes no risk. Too high a rake chases away good players by making it hard to make money. That's not necessarily a problem for the casino. Good players do attract business, but they also walk away with money. But too high a rake also makes the losers broke too quickly. That's also not necessrily a problem, because it means the casino gets more of the losers losses relative to what the good players take.

One way to manipulate things is bad beat jackpots which subsidize the bad players at the expense of the good. High blinds relative to limits also increases the luck factor. Other promotions and rewards can direct money to bad players, away from good.

There is a fundamental difference in economics between casinos and Poker cardrooms without casino games. Casinos love losers, they comp them, lend them money, forgive debts and treat them like kings. Cardrooms love good players. They usually don't comp, but they lend money, forgive debts and treat them well.