ICM and Chip Mapping (part 1)
Posted by Apoth
Video Player is loading.
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Resume Video
Start from Beginning
Watch Video
10
Loading 16 Comments...
great vid.
whats ur opinion on ICM tools like ICMizer, specifically the quiz they offer?
Hi great video, great intro for a cash PLO player here. One question: is there a PLO ICM calculator? I'm sure PLO tournaments are super interesting ICM-wise and certainly misplayed by a lot of people.
MaMaMat I don't have one but that doesnt mean one doesn't exist.
@David ICM tools are only as good as the model. So, in spots where ICM is a very good approximation of reality, those tools are good. In spots where ICM is a mediocre assessment of reality (but people still rely on it) the tools provide mediocre results.
hey man, nice introduction to icm and chip mapping sounds very interesting as alternative. You mention briefly all the other factors that need to be considered in certain spots that ICM can not consider, like blinds passing, edge advantage etc.
I may or may not have one more factor which im not sure about though: So when we model any ICM situation for a normal tournament, it is usually in the money, specifically usually at the final table although one could also calculate situations before I know. But ICM always assumes we are there already, like in the money and so it values for example a payjump from 6th to 5th place highly. However when we look at tournaments realistically, in order to get to the final table we are spending quite a few buy ins on average "prior" (again on average, ofc one can single table first mtt and win) to that. Is it not important to consider the average losses prior to the final table or is that irrelevant? Because intuitively, 5th place 6th place etc in mtts dont feel too great, simply because the cashes get eaten up by losses more quickly again as opposed to 1st-3rd place cashes. So to me it would seem like 8th-4th place finishes are overvalued in the ICM model which leads to too tight actions even if we type in the correct ranges in an icm calculator or if everyone acts according to icm.
It's not entirely clear to me what you're asking.
I think in live events there's some extra value to finishing first and to a lesser extent 2nd and 3rd due to endorsements and other opportunities.
(Like the payout for the main at the top 5 might be something like: 8.5M, 5M, 3.8M, 2.7M, 1.9M) but really, first has a good chance of being signed to a deal with stars or some other company and probably gets into good private games because they're the world champion etc. So in that sense 5th is over valued by ICM compared to first. But that's not actually because ICM is wrong, it's because you've simply put in the wrong number for first prize. Really, first prize is worth the 8.5M cash immediately as well as the (lets just pick a random number) 2M in extra EV. Perhaps 2nd gets 750k in extra EV for being "runner up to the world champion" and 3rd gets 250k extra because they are playing on the final day (so maybe some company pays them to advertise a patch on the final table).
Now really, you've just got adjusted payouts of 11M, 5.75M. 4.05M, 2.7M, 1.9M. If you plugged in those numbers you'd get a more accurate representation of correct play.
In tournaments in general (online especially) where that's not generally the case, I don't see how 4th-8th are undervalued. The fact that you've lost money prior to the position you're in has to effect on the relative current payouts.
If that was the case then the ICM solution would be different for someone at the end of a downswing than for someone coming off a win.
"If that was the case then the ICM solution would be different for someone at the end of a downswing than for someone coming off a win."
I didnt mean it like that. There is always an unknown "expected average loss before final tabling" is what I am thinking. Of course i dont mean it in a results oriented way like someone is on a downswing of 9k now he needs to only care about the 10k first place because every other place he is still down. I just mean that there realistically is always some loss before the final table on average. Whatever that average loss is, I was asking if that couldnt effect the ICM calculations as they dont consider it. But I guess your answer is still no so nevermind.
Ah, sorry i misunderstood. My answer is still no. I don't think such a thing would apply.
Really loved this video! Have been studying ICM a lot lately and am very frustrated with the lacking feature sets of all ICM tools currently in existence. My go to right now is still ICMizer, but even that is majorly lacking, I don't even think mixed strategies can be implemented in that calculator (could be wrong). It's pretty motivating that there is so much more room for refinement in our models even in 2016.
HRC is already beginning to implement mixed strategies
hey, finished the 2nd part of the video. What's the video you refering to at the end: "Luke" or something?
are there nash equilibriums for ICM ($EV)? That is, classical nash equilibriums are based on maximing chip EV, not $EV and so I'm curious if there is software that produces $EV nash equilibriums...
when you procrastinate your lin alg homework by watching a run it once video and then linear algebra is a main focus in the rio video
That's a bad fkn beat. My condolences
Hey Apotheosis,
When can you expect to have a negative chip ev? what number is it to have a negative Chip EV? What would be an example.
also where are you getting the input numbers? what is the input numbers from? Also when you multiply the CEV by the frequency when is it a good number or bad number?.....How do you know?
Hi there,
Great video!
Could you explain a bit how Turbo affect our push fold range? We have less chance to have big hands which dominate our opponents' range. And how should we adjust our strategy in Turbo Games?
Thanks
Be the first to add a comment
You must upgrade your account to leave a comment.