2007PBWC NLH FinalTable - Bot mistake? or misfortune?

[RIIT]

[ QUOTE ]
[ QUOTE ]
I agree with you if this had been a regular cash game where pushing A9o preflop after a reraise from the bb seems way too risky - as I said already, the bots ROR was already at 46% against all hands.

[/ QUOTE ]

Effectively, it's a cash game. ROR doesn't sound very important unless you think this bot's overall skill edge is huge, which I doubt it would be against other decent bots. If I know it'll call with enough hands such that A9o is +cEV against that range, then shove away.

[ QUOTE ]
I'm totally speculating as to what Guaran would have done but I think it's safe to say he'd have folded his entire range except for AA - QQ.

[/ QUOTE ]

Yeah... that's the easy part. The hard part is his opening range, but there's no way to answer the question without it.

Also, it's so cute that no one even attempted to buy the 1k seat.

[/ QUOTE ]

Slim,

Yes I guess it is like a cash game (but not a typical one), in the case where everybody has their entire roll at the table.

The bots risk of ruin was 46% (all hands considered) which is not that high for a single isolated preflop bluff, but quite high if the bot intends to suffer this risk repeatedly (and we'd need to hear from the bot owner to get further clarification here).

The players can't really "buy" an additional stack as such. In the case where there would have been a lone entrant in a $1k event, there would have been 10 minutes time for other players to have the chance to call the first player and enter. So a lone entrant can't just walk up to the cashier seconds before the entry window closes and spend the $100 event fee to get another final table stack. They would have actually had to fork over a total of $1100 ($1k entry fee + $100 event fee) and then they'd have to wait 10 minutes for other potential players to decide whether or not to enter.

RIIT

[Slim Pickens]

OK, I'm going to try again to explain how to analyze this play... Because of the winner-take-all payout structure, this game needs to be analyzed as a cash game. This means evaluating the expected value of each play in terms of straight chip equity.

Risk of ruin is not a useful metric here, and the number you keep stating is not the risk of ruin from making that raise. ROR is not useful because it is redundant information once chip equity analysis is done correctly. 46% is not necessarily the risk of ruin, and in fact if the ROR actually turned out to be 46%, I can promise you the play is way -EV. Where is that number even coming from? A9o vs. a random hand is close to 60/40.

Anyone who attends a STTF event and doesn't at some point wager at least $1000 on something (probably a random event or a contest in which the bettor has no useful knowledge) is considered a major nit, and that's amongst a small group of specialists. This is why I think it's funny that the 1k event didn't even go out of the entire player pool.

[RIIT]

[ QUOTE ]
OK, I'm going to try again to explain how to analyze this play... Because of the winner-take-all payout structure, this game needs to be analyzed as a cash game. This means evaluating the expected value of each play in terms of straight chip equity.

Risk of ruin is not a useful metric here, and the number you keep stating is not the risk of ruin from making that raise. ROR is not useful because it is redundant information once chip equity analysis is done correctly. 46% is not necessarily the risk of ruin, and in fact if the ROR actually turned out to be 46%, I can promise you the play is way -EV. Where is that number even coming from? A9o vs. a random hand is close to 60/40.

Anyone who attends a STTF event and doesn't at some point wager at least $1000 on something (probably a random event or a contest in which the bettor has no useful knowledge) is considered a major nit, and that's amongst a small group of specialists. This is why I think it's funny that the 1k event didn't even go out of the entire player pool.

[/ QUOTE ]

Slim,

I mangled the ROR calculation (entered the bots stack incorrectly the first time); the actual ROR is 57.37%, which is severe by any standard.

The ROR calc I'm using is the statistical proposition of doubling your current stack before losing it - where the stack is what the bot had before pushing. Note that the ROR is close but not identical to the bots losing chances for the hand of 56.69%

RIIT

[DevinLake]

what is ROR? I know it means risk of ruin, but I've never seen it used in poker outside of bankroll management.

This is a pretty simple situation to look at. You have said the bot will probably not call with less than QQ. So, with that calling range, we can easily figure out what he needs to be min 3betting here to make this a +cEV shove.

[RIIT]

[ QUOTE ]
what is ROR? I know it means risk of ruin, but I've never seen it used in poker outside of bankroll management.

This is a pretty simple situation to look at. You have said the bot will probably not call with less than QQ. So, with that calling range, we can easily figure out what he needs to be min 3betting here to make this a +cEV shove.

[/ QUOTE ]

Devin,

In the most typical usage it's the statistical probability of losing X before winning X. And X can be anything you chooose (bankroll, stack, etc.)

For the case in point here, I consider X to be AsbakAlpha's stack before the push which was 952.50; but if this was a cash game then we'd probably use his entire roll instead.

I said that I believe that Guaran (not the bot) would not have called with anything less than QQ (just to clarify here).

RIIT

[Slim Pickens]

Can you elaborate on exactly what you're doing to calculate the bot's losing chances for the hand of 56.69%?

[RIIT]

[ QUOTE ]
Can you elaborate on exactly what you're doing to calculate the bot's losing chances for the hand of 56.69%?

[/ QUOTE ]

Slim,

That is the statistical chance of losing to the range of all possible hands. The number gets worse if Guarans preflop raising range is smaller (which we can only guess at).

RIIT

[pokerdude@ub]

[ QUOTE ]
[ QUOTE ]
Can you elaborate on exactly what you're doing to calculate the bot's losing chances for the hand of 56.69%?

[/ QUOTE ]

Slim,

That is the statistical chance of losing to the range of all possible hands. The number gets worse if Guarans preflop raising range is smaller (which we can only guess at).

RIIT

[/ QUOTE ]

Oh Ray....cough RIIT [img]/images/graemlins/cool.gif[/img]

People don't call based on random hands (unlike the way Winholdem thinks). Call $900 on a 1/2 blind is insane (unless you have AA).

Making that bet is worst.
With one guy to act he has AA 1 in 216 times. If two guys are left to act that is 1 in 108 times (between the two of them).

HUGE -EV play.

Despite what your iterator tells you.

[Tantalus747]

[ QUOTE ]
I agree with you if this had been a regular cash game where pushing A9o preflop after a reraise from the bb seems way too risky - as I said already, the bots ROR was already at 46% against all hands.

[/ QUOTE ]

You keep bringing that number up, but villain doesn't have a random hand, he has a reraising hand from out of position. I would suggest a study of Bayes' Theorem if you have not.

I started to say even pushing A-Q there would be horrific but then caught myself; there's really no hand that makes that overbet look sane.

Edit to add, maybe you're putting Gauran on a much wider range than everyone else here? That's the only way I can read this.

Also, good players who think they have an edge in that game would toss even QQ to a overbet that size (absent a good read obv.)

[RIIT]

[ QUOTE ]
[ QUOTE ]
I agree with you if this had been a regular cash game where pushing A9o preflop after a reraise from the bb seems way too risky - as I said already, the bots ROR was already at 46% against all hands.

[/ QUOTE ]

You keep bringing that number up, but villain doesn't have a random hand, he has a reraising hand from out of position. I would suggest a study of Bayes' Theorem if you have not.

I started to say even pushing A-Q there would be horrific but then caught myself; there's really no hand that makes that overbet look sane.

[/ QUOTE ]

Tantalus,

The 46% was my error. The correct ROR number is 57+%. Yes I agree that the push was a mistake (some of my previous comments were due to the incorrect 46% number).

RIIT