ICM calculations

I was just listening to the recent show and heard you(the wizard) say that you were not familiar with ICM calculations. Prior to the UIGEA, I specialized in single table tournaments which meant a lot of late game pay. In tournaments, chips don't map directly to dollars. You can get all the chips in a tournament and you don't get all the dollars. Chips at the top of your stack are worth less than those at the bottom. This effect is particularly pronounced if you're close to the bubble. It's been a while, but if I remember correctly, in the old party structure(50/30/20), if you had one of 5 equivalent stacks on the bubble and doubled your stack, you only increased your value by about .6. The math for this was done by a 2+2er named Eastbay and included in a lesson in his early commercial ICM calculator, SnG Power Tools.

Among player who understood this, this could lead to a large skewing of strategy, known as "The Gap", where large stacks could risk little value and play very aggressively on the bubble, especially if one of the stacks was really short, enhancing the effect. Of course, if the other players didn't know this and were obviously playing too loosely, you had to tighten up, your value coming from letting the overly loose player bust out. At times the big stack would preserve the bubble by not forcing the short stack all-in in the big blind.

Of course, some of this is intuitive, but ICM gives people a way to handle it with precision, and precision was worth a lot to Sit n Go players who made hundreds of these decisions a day.

The actual algorithm is modeled after a lottery if I remember right, but I only glanced over it once.

Hope this was interesting to someone,
John

socks,

Very interesting to me, thanks for posting it! Listened to the same podcast, wanted the same explanation.

I am left scratching my head at this entire situation. The original article the wizard linked too mentioned this was properly priced according to icm. The thread the wizard started mentions icm a few posts in. For the life of me I can't explain why he would go on a radio show days later and make a statement that was not only false but lacked any basis in logic. Even if you don't get icm you should understand tournament chips are not directly translatable to money in proportion, especially if the tournament is not winner take all.

The op here is very kind and helpful but I don't get the lack of diligence by others.

I mean some poor sap could have dropped ten dimes on this.

I admit it: I'm behind on my podcasts.

As a poker player, the OP is quite interesting. But (he asks sheepishly...)
What does ICM stand for?

Quote: DJTeddyBear

I admit it: I'm behind on my podcasts.

As a poker player, the OP is quite interesting. But (he asks sheepishly...)
What does ICM stand for?

"Independent Chip Model". It makes the assumption that each chip helps improve your chances of finishing 1st place in the tournament equally. So if you have 7000 of the 10000 chips in play, then ICM will say you have a 70% chance to finish in first place. The math gets a lot more fuzzy when you need to calculate their chances of finishing 2nd, 3rd, etc.

Very, very interesting words. I can recall many occasions in TH-NL tourneys (albeit at "micro stakes" under $5) where being on the bubble, and not a blind, fold anyhing... even Rockets.. Time, ultimately weighs favor. However, this posting appears to be aimed at single table "Sit and Go's", where patience is usually well rewarded with a cash-out.

Here is it Teddy

http://www.thepokerbank.com/strategy/tournament/icm/how/

As useless as most other formula's. Does not take into account how much an edge you do or do not have over the other 2 players
in the game. Whether you call or do not call the aggressive player has more to do with the unknown facts that the know facts.

If the other 2 are better players, I might want to call and hope to be chip leader to improve my chance of winning 1st, plus guarantee me a payday. If I am much better than the other 2 players, I would never call the all in player. This is another formula for people to
justify their play, win or lose. I only had 26 Big Blinds. I was 62% favorite over a random hand. ETC.

Note in the example the all in player is aggressive and will go all in with any 2 cards. How aggressive passive are the other 2 players?
Will one of them call this fool in the next few hands, guaranteeing you a cash position. Since the HERO knows this player is very aggressive, why does he not share info about the other 2 players.

What is the hero's table image ? So many unanswered questions. ICM is good when playing low limit against donkeys. But then, why bother ?

Quote: Buzzard

As useless as most other formula's. Does not take into account how much an edge you do or do not have over the other 2 players
in the game. Whether you call or do not call the aggressive player has more to do with the unknown facts that the know facts.

If the other 2 are better players, I might want to call and hope to be chip leader to improve my chance of winning 1st, plus guarantee me a payday. If I am much better than the other 2 players, I would never call the all in player. This is another formula for people to
justify their play, win or lose. I only had 26 Big Blinds. I was 62% favorite over a random hand. ETC.
...

The default chip valuation model, directly mapping chip value to dollar value, doesn't solve all these problems either. Still, you need some model to value chips when analyzing decisions and most good players believe this one to be better than that default.

Also, this model is known today because good (winning) players valued it and discussed it early on, players who usually didn't tolerate bad beat stories.

Quote: randomperson

I am left scratching my head at this entire situation. The original article the wizard linked too mentioned this was properly priced according to icm. The thread the wizard started mentions icm a few posts in.

I didn't realize there was another thread. I barely followed poker when I played, and when I stop by here, I usually look at "Math", and "Game Inventors Corner". I probably would've just commented there if I'd known.