ICM-based decision making at the final table?

SunOfBeach · #1 02-07-2006, 04:30 PM

Hi,

I'm finishing up a book on a game theoretic analysis of the endgame of MTT's (PS it will be apples and oranges above kill phil, etc in terms of mathematical sophistication... but admittedly geared to a much different audience).

I have a question to anyone who'd be willing to help me out: does anyone actually consider icm-type analysis (ie prize structure implications) when they're at the final table? or is everyone basically just shooting for first, since that's where the real money is. would anything be gained by systematically analyzing these types of things, or is it so intuitively obvious to shoot for first that it becomes boring to look at the math of it?

thanks

hockey coach · #2 02-07-2006, 05:05 PM

I'd love to see the math of it - "play for first" is an oversimplification for "play to maximize EV". I attempt to account for this at the FT (and other times) but I'm sure I've oversimplified as well and or am just flat out wrong at times. The stochastic nature of the game is very complicated (at least to someone who last studied or used stochastic processes during the Carter administration) and I'd love to get some help.

Regards,
HC

zoobird · #3 02-07-2006, 05:33 PM

Sounds interesting to me. I play mostly MTT sit-n-go's, where the payout structure is somewhere in between a big MTT and and a single table tournament, and my results seem to have improved since I stopped 'shooting for first' quite as much. Having some mathematical tools to determine exactly how to gear strategy to prize structure would be great.

Mvcode3 · #4 02-07-2006, 05:51 PM

Defintily worth taking a look at, how do you factor in skill levels with ICM--Ie im at a final table with Ivey, Hellmuth, and brunson with the chiplead....are my chips worth "less" then theirs in terms of payout?

esad · #5 02-07-2006, 05:53 PM

I think it depends on how you look at it. I am always trying to win every MTT I enter. But, that doesn't change my style of play. Sometimes it's a mind game thing--people playing for first will either take too many chances to be high in the chip count or will tighten up for fear of getting knocked out.

Sorry, but I just don't see any point of engaging in any competitive endeavor unless you are trying to win.

NHFunkii · #6 02-07-2006, 05:54 PM

my intuition tells me that it's close enough to irrelevant in most cases, but when it gets to the last few spots and there's a shortstack it's worth taking a look at.
9 handed though I don't think it's terribly significant. but I could be wrong.

SunOfBeach · #7 02-08-2006, 01:08 AM

ok, im going to post more in the coming weeks on some of this type of stuff. thanks for your reponses,

peter chabot

Proofrock · #8 02-08-2006, 01:12 AM

it's funny you should post this, because i've been intending to write a little something about this within the next week or so.

SumZero · #9 02-08-2006, 06:18 AM

I think you'd be crazy not to consider ICM when you are at the final table. Play for first is an oversimplification that is often right but not always.

fiskebent · #10 02-08-2006, 09:55 AM

ICM can support decisions at the FT that some may find intuitively obvious and others may not. For instance if there are 7 short stacks and 1 big stack at the FT with you, it's almost always -$EV to go up against the big stack for your entire stack no matter how large your edge is. This is one of the widely known scenarios where it's the right decision mathematically to fold AA pre-flop.

An add-on to ICM could be the non-linearity in the value of dollars when the prizes are large compared to your bankroll / general economy. You can say that the 'utility' you get from winning $1 mio. isn't always 10 times as large as the utility you'd get from winning $100,000. If I win $100,000 maybe I can pay off some debt and make my life easier for myself and my family for the next 30 years. A $100,000 increase in prize money from $900,000 to $1,000,000 might not give me nearly as much value for money. Maybe I'd blow those last $100,000 on a new car or something.

Another scenario where there isn't a linear relationship between dollars and utility could be one where I owe a goon $800,000 and he'll break my kneecaps if I don't pay in 2 days. That would make the prizes below $800,000 nearly worthless to me and I might take more risks to win a large prize.

If the relationship between dollars and utility isn't linear, your should go for maximising utility rather than dollars.

If the prize money are insignificant compared to your general economy, there'll probably be a linear relationship between dollars and utility and you can just play to maximise $EV.