ICM Quantified - First Set of Results

[rvg72]

[ QUOTE ]
[ QUOTE ]
this [censored] rules.

keep it up.

c

[/ QUOTE ]

I agree! I really feel like we (you) are achieving something here.

edit: maybe a stupid question, but could this be caused by calling ranges? A (very) short stack will often be called by at least one, and often two players if he goes all-in. A big stack on the other hand has more FE.

[/ QUOTE ]

Calling ranges / Fold Equity are definately key reasons for this but they are also reflected in your low ICM value - it is low because you don't have as much Fold Equity. What this shows is that your low stack is even worse than you thought.

Now this doesn't mean that you push any 2 from any position at any blind size if you are a low stack on the bubble. You are not going to steal any more often or win the all-in any more often with this new knowledge... but... when you are called and lose you don't lose as much since your $ev was lower anyways and if you succeed in the steal or double up then you make more real $ev than you would have expected.

rvg

[rvg72]

[ QUOTE ]
do u want more HH? i'm sure there are people here who will send u thousands of HH to help this analysis.

[/ QUOTE ]

YES - I have had a couple of donations which has helped out. I need full hand histories - from first hand to last hand. Any help would be appreciated and all results will be posted here.

rvg

[ZeroPointMachine]

[ QUOTE ]
[ QUOTE ]
do u want more HH? i'm sure there are people here who will send u thousands of HH to help this analysis.

[/ QUOTE ]

YES - I have had a couple of donations which has helped out. I need full hand histories - from first hand to last hand. Any help would be appreciated and all results will be posted here.

rvg

[/ QUOTE ]

The only full HH I have are 1st and 2nd place finishes. If I send you all of my HH can you parse out the ones that will work?

[zabt]

So, you mean that ICM more accurately values a small stack at higher blinds than at low blinds? Or, put another way, it is worse (relative to ICM) to have a short stack early than a short stack late? I suppose some of this is because when you have a short stack late, you're already late into the SNG, closer to the end, less random poker to be played, less time for error to accumulate so to speak. Or, maybe it's just because really bad players, the ones who are worse than random, are often short stacked early and their results are skewing the data. Come to think of it, the donks are more often out early so their datapoints will not be present as often in the higher blind levels. It might be that the good players are the ones that accumulate the large stacks and thus the reason that big stacks perform better than ICM predicts is that better players are playing the big stacks. And, the bad players are the ones that find themselves with short stacks which is why short stacks come up short of ICM prediction.

I guess you'll need to break down the data points by player ROI (or some other measure of skill) to see if short-stack worse than ICM/large stack better than ICM holds across player abilities.

[Phil Van Sexton]

[ QUOTE ]
Very cool. And, I suspect not a big surprise. I'm very interested in seeing the by position results.

[/ QUOTE ]

Yes, well done. I made a token effort at this a few months ago, but got bogged with real work before concluding anything.

This was my chart by position (also by blind level). ICM overvalues the play in BB (duh), and undervalues the button in my small sample.

I think we could get some traction here if you published your database and/or code. If you could collect hand histories, remove the names, and publish a database....then other people could run their own analysis, and double check each other's results.

[rvg72]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
do u want more HH? i'm sure there are people here who will send u thousands of HH to help this analysis.

[/ QUOTE ]

YES - I have had a couple of donations which has helped out. I need full hand histories - from first hand to last hand. Any help would be appreciated and all results will be posted here.

rvg

[/ QUOTE ]

The only full HH I have are 1st and 2nd place finishes. If I send you all of my HH can you parse out the ones that will work?

[/ QUOTE ]

My app will disregard any SNG's that don't have final results for all 10 players so that won't be an issue - my only concern would be that the results might be biased. I would prefer watched hand histories where 10 semi-random people took part but that being said, I would gladly take anything you give me. I will add a new table in the DB that identifies if it is a "Biased" dataset so I could seperate the results if need be.

Thanks,

rvg

[good2cu]

wow. This could have quite a big impact on tournament poker, espically if the results stay the same through a larger sample.

[AliasMrJones]

Interesting. The sample size isn't very big. Getting lots of tourneys from lots of other people would be great. This could be paradigm-altering research -- congrats! Let us know how to donate HH's.

[rvg72]

[ QUOTE ]
I think we could get some traction here if you published your database and/or code. If you could collect hand histories, remove the names, and publish a database....then other people could run their own analysis, and double check each other's results.

[/ QUOTE ]

I would be very interested in publishing the database. For starters, here are the fields I am collecting for each datapoint (a datapoint is 1 hand for 1 player):

Stack Size
Number of BB's
Level
ICM Value
Actual Result
Buy in
Players Left

So, if you played 64 hands then you would have 64 data points in the file.

What should I add?

Zabt mentioned adding a measure of skill - can't do ROI since I'm not tracking player results across more than the one SNG. Saw Flop % could be added for each data point so a player who saw the flop 10 out of 50 times would have 20% as their Saw Flop % for every Data Point from that SNG - it would not carry over to the next time they play or any previous SNG's they played though. I think it has some value after the first 30 hands or so - bad players see a lot of flops, good ones don't.

I will definately be adding distance to BB.

Any other ideas?

rvg

[rvg72]

[ QUOTE ]
Interesting. The sample size isn't very big. Getting lots of tourneys from lots of other people would be great. This could be paradigm-altering research -- congrats! Let us know how to donate HH's.

[/ QUOTE ]

Thanks - PM me and I will provide an e-mail address to send zipped up histories to.

Thanks for helping out,

rvg