4-of-a-Kind
Join Date: Oct 2005
Location: St. Louis
Posts: 3,654
|
|
ICM might have some weaknesses. For example, it assumes that everyone's chance of finishing 1st is proportional to the size of their stack. Sometimes this isn't the case, because one player is likely to have a lot of good situations to keep picking up pots. Meanwhile another player is unlikely to win any pots unless he gets dealt a monster. The situation in taipan's hand is a good example, because the chip leader is probably going to keep pwning the bubble until one of the short stacks calls him, and taipan will keep getting blinded down until that happens (or he gets AA and calls). So, while ICM says the villain has about a 40% chance to win and the hero has about a 30% chance to win, in reality the villain's chances are better than that and taipan's chances are lower.
How does this impact our play in an actual SnG? Well, if we're the big stack and we're pwning a bubble, we should be less likely to call a push from a short stack, because SnGPT is going to under-estimate the equity we'll have if we fold (because it doesn't count the chips we'll pick up by continuing to pwn the bubble). Conversely, if we're the second stack and we're getting pwned, SnGPT is going to over-estimate the equity of a fold. Therefore, we should be calling with more hands than SnGPT says we should call with.
How big of a difference does this make? I don't have a clue, but my feeling is that calling with a hand like 99 (which is borderline according to SnGPT) might be good because of the above considerations, but AK is such a horrible call according to SnGPT that it's probably still bad.
This is something I've been meaning to post for awhile but I just never got around to articulating my thoughts about it.
|