Understanding ICM analysis in Coolers

[GaF]

Lets do it :ok

NLHE STTs are pretty much a "solved" game. Coolers are just a slight variation of the standard STTs - someone who has live access to ICM information in coolers should crush these :unsure

if you can understand all that stuff. GaF any chance you can do a step-by-step guide to understanding what those fancy tables mean - a real idiots guide please?. I know that it is a ball-ache, but it would be very helpful Cheers Damo

Well said that man. :clap:clap:clap

Eeeek!!! I dont mind doing whatever I can to try and help PLrs get their heads around this, however I lack a bit of confidence when it comes to "teaching" it to others :unsure I dont think my understanding/knowledge of it is 100% (hence I panic every time I see Slapdash in the thread :tongue2) and that makes it tough to explain with clarity and precision - I am certain that there will be better sources of information out there than what I can come up with :unsure Though maybe not the in specific relation to coolers. So what I'd propose is finding external sources to understand the common aspects and then I will try and put some explanation into how I'm applying that to coolers - how does that sound? I'll look around for information to reference over the weekend - the way I see it is that there are 3 steps to ensure you understand pretty well (and each is a prerequisite to the next): 1) Expected Value (ev) 2) STT ICM 3) ICM theory applied to coolers I suggest we work through 1 and 2 together (ensuring everyone understands 1 before moving on to 2) from external sources and then work through 3 creating our own reference material. Does that sound reasonable?

is this a good way to do it? pick a hand and post an ICM table like you have used before with tags showing us what each column means and a basic if it is above this figure it is positive to play, if it is below that figure it is negative to play then post another table showing the next hand and do the same again then post a third table showing the next hand showing the same again that way we can see small changes to the tables and the ICM calcs. This I think would aid us to see a series of hands evolve and show the steps on the tables to give us a better understanding of what the numbers are showing. What I am particularly interested in is how we could use the info in a 'live' game and make a snap decision based on findings. I think that post game analysis with pokerstove and opinions is vital, but something that aids you in your 15 sec decision time in a hand is critical. Would the ICM calculator do that for you? I am prepared to wait and let the maths guru's among you sit and discuss first rather than expecting you to do it all this weekend:ok It is a tall order and most appreciated, so take you time and do the basics first (like posting the table and telling us what each column and figures are showing!) Damo

I'm not aware of any tools that will help you with cooler ICM analysis in 15 seconds - with the examples I've posted you can see that the time between the first post and last post is about 30 minutes - I've no chance of replicating it within 15 seconds. I might be able to create an AHK script that could use the tools to do it that quickly - not on Boss (I cannot get AHK to read live Hand information) but maybe on Pokerstars - however (without having checked) I'm pretty sure it would be against Stars T&C so wouldnt want to waste my time doing it. More than happy to go step by step through the process I do and explain what I can - however anyone who wishes to interpret the results properly must understand at least basically, concepts of ev and ICM first.

excellent:ok - please can we start with "what all these numbers mean on the big tables I keep posting in my analysis posts" Cheers:ok Damo

Lets start a seperate thread for this :ok I'll copy some posts over ....

[GaF]

Re: Understanding ICM analysis in Coolers Required tools on your PC: Pokerstove Website - http://www.icmpoker.com/Calculator.aspx Spreadsheet

[GaF]

Re: Understanding ICM analysis in Coolers Quote:

Originally Posted by robilaruk excellent:ok - please can we start with "what all these numbers mean on the big tables I keep posting in my analysis posts" Cheers:ok Damo

I presume this is the kind of "big table" you are talking about? It's a screenprint from the ICM Calculator web site. I'll go through the columns in next post.....

[GaF]

Re: Understanding ICM analysis in Coolers The important columns are the first column and the final column. First column is the chip stack of each player - you enter the chip stacks here. The final column is the proportion of a prize that the player can expect to win given their current chip stack. In coolers, the top 5 players win 100% of a prize, the bottom 5 players win 0% of a prize. This number will always therefore be between zero and 100. It will be 0 when you are eliminated and have no chips left. It will be 100 when you still have chips and there are 5 players left. At the start, when all players have the same number of chips it will be 50. So in this specific example, within the constraints of the weaknesses of ICM analysis, player 1 has a 95% chance of winning the prize,whilst player 2 only has a 9.5% shot at a prize. This can be converted into "expectation" - if it's a €5 +50c game with a €10 prize, then Player 1 effectively holds €9.50 in the tourney whilst player 2 effectively holds €0.95 equity in the tourney. These players should play where they feel they can increase the monetary value of their expectation in the tourney (against the monetary expectation if they fold).

[GaF]

Re: Understanding ICM analysis in Coolers The middle columns aren't that important to understand (basically the working out), which is why I've put this in a seperate post. You must complete the prize structure in the headings - for coolers it is 5 places with the same prize - you can put any number you like in there (so long as it's the same in all 5 spots) - I prefer to put 100 in all the time as it gives nice round numbers in the output (percentage of a prize). The working out (if you have "place %" selected or with either option if you have used the value 100 for the prizes) is the percentage chance each player has of finishing in that position. So player 1 is estimated to have a 27.3% chance of finishing 1st, a 23.7% chance of finishing 2nd, a 19.6% chance of finishing 3rd, a 14.8% chance of finishing 4th, a 9.6% chance of finishing 5th and a 4.9% chance of missing out on the top 5 spots - add those numbers up and it should come to 100%, so covering all eventualities. If you look vertically instead of horizontally, then you can see the likelyhood of each player finishing within a set position, so for example, the winner is 27.3% likely to be player 1, 1% likely to be player 2, 22.2% likely to be player 3, 17.9% likely to be player 4, 11.9% likely to be player 5, 8.5% likely to be player 6 and 11% likely to be player 8 - that should add up to 100% to account for all posibilities about who the winner may be.....

[GaF]

Re: Understanding ICM analysis in Coolers Are you still with me, and any questions on this aspect?

[daftpegasus]

Re: Understanding ICM analysis in Coolers

Are you still with me' date=' and any questions on this aspect?[/quote'] Yes - What does ICM stand for?:unsure

[GaF]

Re: Understanding ICM analysis in Coolers ICM = Independant Chip Model (confess I had to look it up :$)

[robilaruk]

Re: Understanding ICM analysis in Coolers

Are you still with me' date=' and any questions on this aspect?[/quote'] EXCELLENT - this is just what I was asking for:clap:clap:clap Thank you:ok Damo looking forward to the next bit and saving his questions until the end

[GaF]

Re: Understanding ICM analysis in Coolers If you have questions on this - best to ask them now I think :ok

[robilaruk]

Re: Understanding ICM analysis in Coolers BTW can this be stickied for a bit please? Ta Damo

[robilaruk]

Re: Understanding ICM analysis in Coolers

If you have questions on this - best to ask them now I think :ok

the questions are more about game play than the software/figs - as in who best to target and why And I have a feeling the EV bit will answer the question of when to get involved (i.e if +ev over tourney % then play, otherwise fold) Damo ps hang on - BRB, poorly child just awoken, a couple of stories and she will be asleep again! :)

[GaF]

Re: Understanding ICM analysis in Coolers Ok - I'll move on then :ok So given chip stacks, we can now pretty accurately estimate the prize equity someone holds before or after a hand. During a hand we can use a "decision tree" to determine which action gives us the highest prize expectation.... Lets find a hand.....

[GaF]

Re: Understanding ICM analysis in Coolers Lets use the hand here - 20081114 - Cooler Analysis - Hand 34

***** Hand 1320762584 ***** 150.00/300.00 Texas Hold'em (No Limit ) - 14 November 2008 19:34:40 Cooler (Real /Tournament ) Seat 2: jigen74 (2944.00) Seat 3: XenderAlli (2220.00) Seat 4: RayPenber (858.00) Seat 5: aALjane (985.00) Seat 6: PL-Aldric (1555.00) Seat 7: pl---GaF (1305.00) Seat 9: albi70 (3778.00) Seat 10: nike123 (1355.00) PL-Aldric post SB 150.00 pl---GaF post BB 300.00 ** Deal ** jigen74 [N/A, N/A] XenderAlli [N/A, N/A] RayPenber [N/A, N/A] aALjane [N/A, N/A] PL-Aldric [N/A, N/A] pl---GaF [:2s:, :2d:] albi70 [N/A, N/A] nike123 [N/A, N/A] *** Bet Round 1 *** albi70 Fold nike123 Fold jigen74 Fold XenderAlli Fold RayPenber Fold aALjane All-in 985.00 PL-Aldric Fold pl---GaF ????????

We are last to act, so that simplifies the situation enormously. First thing we need to do is be aware of all the possibilities. In this case, someone has shoved into us, so we only have two options - we can call or we can fold. Which should we do? We should do the one that leaves us with the most tournament equity! If we fold, it's easy - the person who shoves wins the hand and we know exactly how many chips everyone will have - punch these chips into the "ICM Calculator" and we're left with the equity below: We are left with 1005 chips which means we have 44.6% prize equity - expressed differently - if this were a €5+50c cooler with a €10 prize, then our seat in the tournament if we fold is worth €4.46. Now we need to see how €4.46 we have for folding compares with what we can have if we call........

[GaF]

Re: Understanding ICM analysis in Coolers Now it's a bit more difficult to work out our equity if we call. Why? Because we dont know what cards our opponent holds and even if we did know what our opponents cards were, we dont know which hand will win the showdown, and we know that we are going to showdown...... First thing to do is to estimate our chances of winning the hand this is done with Pokerstove and by putting in a range that we believe our opponent could shove with in this situation. If we believe our opponent will shove with any two cards, then our chances are: If we believe that our opponent will shove with a top 50% range of cards, then our equity is: To start off with I will examine the maths behind believing our opponent shoves with quite literally any two cards - so look at the upper pokerstove screenshot and ignore the lower one. This gives us 49.67% equity in the hand (in tournament chips, not tournament prizes). However it's a touch more complicated than that.... Think again in terms of a "decision tree" - what are all of the options if we call? - We can call and win the hand - We can call and lose the hand - we can call and split the pot We need to identify the chances of each of these 3 scenarios occuring - so we need to ignore the "headline" numbers in pokerstove and look a the "output" box. We can see that if we call, we are 48.72% likely to win the hand If we call, our opponent is 49.39% likely to win the hand The chances of us splitting the pot (and our opponent splitting the pot) is 0.95% + 0.95% = 1.9% As a double check - add the 3 numbers together and it comes to 100% - so we know all possibilities are covered ........

[GaF]

Re: Understanding ICM analysis in Coolers So if we call, we have 3 scenarios to look at, which will give us 3 different chip stacks and 3 different tournament prize equities .... so we need to look at each one in turn. If we call and win, then we have 2440 chips and 80.992% of a prize (which can also be expressed as saying our seat in the tournament is worth €8.10

[GaF]

Re: Understanding ICM analysis in Coolers If we call and lose then we only have 320 chips left, which leaves us with 16.932% tournament equity (with a value of about €1.69) :sad

[GaF]

Re: Understanding ICM analysis in Coolers If we call and split the pot, then we're not that far from where we started - we have 1380 chips which gives us 56.264% tournament equity (with a value of about €5.63)

[robilaruk]

Re: Understanding ICM analysis in Coolers

Lets use the hand here - 20081114 - Cooler Analysis - Hand 34 We are last to act, so that simplifies the situation enormously.

quick request - can you post the ICM at the start of the hand please - am wondering what the stats are and how much they differ before we fold and shover wins. ta Damo

[GaF]

Re: Understanding ICM analysis in Coolers So we know that there are only 4 possibilities: 1) We fold and leave ourselves with a seat worth €4.46 2) We call and win leaving ourselves with a seat worth €8.10 3) We call and lose, leaving ourselves with a seat worth €1.69 4) We call and split the pot leaving ourselves with a seat worth €5.63 So which option do we want? We want to call and win :) Option number two please. Unfortunately it's not as straightforward as that - we only have two choices - we can fold and accept option 1, or we can call and take one of options 2, 3 or 4, but we cannot control the outcome or chose which one. We can however estimate (with the results of post 15) the chances of each option coming up :) 1) If we fold, we are 100% certain of a seat worth €4.46. 2) If we call, we have a 48.72% chance of winning a seat worth €8.10. Multiply these out and the "expectation" is for a seat worth €8.10 x 48.72% = €3.95 3) If we call, we have a 49.39% chance of losing and being left with a seat worth €1.69. The expectation for this seat is therefore €1.69 x 49.39% = €0.83 4) If we call, we have a 1.9% chance of splitting the pot and being left with a seat worth €5.63. The expectation for this seat is €5.63 x 1.9% = €0.11 We can add up the expectation of all 3 results if we call to find our total expectation if we call - so it is €3.95 + €0.83 + €0.11 = €4.89 So we can now compare the value of our only 2 options: A) We can fold and be left with a seat worth €4.46 B) We can call and be left with a seat expected to be worth €4.89 So you now have a choice - would you like €4.46, or would you like €4.89? Most poker players will choose the €4.89 (call) as this is worth 10% more than the alternative. Pheeew ... any questions to here?

[GaF]

Re: Understanding ICM analysis in Coolers

quick request - can you post the ICM at the start of the hand please - am wondering what the stats are and how much they differ before we fold and shover wins. ta Damo

Ok - the equity before any cards are dealt, and before any blinds are paid is as below. Our stack of 1305 chips holds 54.252% tournament equity - so our seat before the hand is worth €5.43. Is that what you meant?

[GaF]

Re: Understanding ICM analysis in Coolers

Ok - the equity before any cards are dealt, and before any blinds are paid is as below. Our stack of 1305 chips holds 54.252% tournament equity - so our seat before the hand is worth €5.43.

Just a couple of observations on this: - The value of our seat before the hand is played, plays no part in our decision making process - it's irelevent - we can not turn back the clock, we cannot change our cards or get a refund on our blinds - we cannot get our €5.43 seat back - our choice is solely a €4.46 seat or a €4.89 seat. - It shows us that we can not make a profit on this hand with any decision we make - the best decision is therefore the one that loses us the least. - Avoid being results oriented - calling and winning gives us an €8.10 seat. That does not mean we profited on the hand, even though the value of our seat has increased from €5.43 to €8.10 - that is solely down to luck/variance - when we call, our seat falls in value to €4.89 - after that there is just some random variation to the value....

[robilaruk]

Re: Understanding ICM analysis in Coolers

Ok - the equity before any cards are dealt, and before any blinds are paid is as below. Our stack of 1305 chips holds 54.252% tournament equity - so our seat before the hand is worth €5.43. Is that what you meant?

yes thank you - was interested in comparing before and after and what happens Obviously the shover's equity increases a lot if you fold, by 13%; Your equity decreases by 10%, but the only other player radically affected by the change is the other short to middle stack in the SB who drops 3% (loss of SB) - it doesn't affect any others significantly, so they dont seem to gain if you fold. (question: if the SB was the biggest stack, would their equity drop by 3% as well) However everyone gains if you call and win and most by about 3-4 % (you gain alot 26%), however I can't understand why player 6 and player 10 who have similar stacks have a wide difference in +equity - player 6 has 0.3% increase yet player 10's increase is over 4% - I realise player 6 has payed the SB and 'lost' it but their overal equity is now Player 6 61.26 Player 10 59.81 yet before the hand they were Player 6 60.96 Player 10 55.68 Thats a massive swing even though player 6 is now 'safe' from the blinds for an orbit, and player 10 is soon to pay them. So ICM does not take upcoming blinds into the equation then. Damo (I think I have answered my own question!)

[GaF]

Re: Understanding ICM analysis in Coolers Correct - ICM doesnt know how big the blinds are, where they are, or take any consideration of them whatsoever (and that is one of the weaknesses of ICM, especially when blinds get very large)

[GaF]

Re: Understanding ICM analysis in Coolers The way the percentages are worked out (the middle columns)..... Each chip is considered a "lottery ticket" with an equal chance of winning. So the chances you see for each player winning (finishing exactly first - not 2nd to 5th) should be exactly equal to the proportion of chips they hold. The chances of finishing 2nd for each player are the proportion of chips that they hold for each circumstance where one of the other players finishes first. The chances of finishing 3rd for each player are the proportion of chips they hold for remaining players, given every combination of the other players finishing 1st and 2nd. etc..... (though it's probably not too important to fully understand how this is calculated - in principal, it's pretty straightforward - just not very well explained :tongue2)

[GaF]

Re: Understanding ICM analysis in Coolers Make sure you understand this Damo - because at the end I'll give you a hand to analyse and ask you to come up with the numbers to check it makes sense :ok

[robilaruk]

Re: Understanding ICM analysis in Coolers

Make sure you understand this Damo - because at the end I'll give you a hand to analyse and ask you to come up with the numbers to check it makes sense :ok

HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA I dunno my first night off in a week and he expects me to do some maths:eyes..... (bloody slavedriver..........:spank:tongue2:spank) Harken? is that my poorly child awake again? oh dear I must run..... honest...... Damo ps reckon I got away with that - he wont notice I was fibbing......:eek pps post it up, no guarentees I will do it tonight or Saturday (we are entertaining!) but I will do it. Most effective way of learning - by doing!

[GaF]

Re: Understanding ICM analysis in Coolers I'll post a hand up in the next day or so (next time I face what looks a tricky decision :ok)

[robilaruk]

Re: Understanding ICM analysis in Coolers

Each chip is considered a "lottery ticket" with an equal chance of winning. etc..... (though it's probably not too important to fully understand how this is calculated -

Dows anyone know how it actually is calculated (the 2-5), not the first? Am wondering if this can be transposed to excel and then other calc's added (blinds/M etc) to keep all the maths in one place rather than many? Just wondering Damo

[GaF]

Re: Understanding ICM analysis in Coolers Yes - the calculations are in the public domain......problem is that by the time you fifth place, it's getting reasonably complicated - it's beyond my excel skills (will need scripting I think :unsure) but imagine it'll be pretty straightforward for someone who can do it :unsure Having said that, not sure what advantage you would get by having available in Excel as opposed to the website :unsure