High stakes tournament crusher Nick Petrangelo views the Independent Chip Model (ICM) differently than most tournament players.
Nick shares his views in his ICM Unmasked course. Here’s a quote from the introduction video:
I think that people have gotten really carried away with the concept of ICM and it being very rigid…but I think there are certainly more cons than there are pros for catering your strategy due to ICM.
What I want to get through [to you] in this series is how important it is to adapt to the strategies of your opponents and to understand the conditions that you’re playing under.
Do your best job of being a good strategy game player [rather than] being a good study-er of this model.
In other words, Nicky P recommends taking in all important factors — some of which are neglected by ICM — in these dynamic situations. This is a notable departure from how some players think about adjusting their strategy when playing for a prize pool and not just chips.
That’s not to say ICM is useless. Let’s look at the “pros” of the model according to Nick.
For starters, ICM is the best model we have and it can be useful as long as we understand the limitations. There are currently no better alternative models.
ICM has two major use-cases:
Firstly, it is great for assigning values to chip stacks so that chops can be made on final tables. Calculating the dollar value of tournament stacks is difficult, so having a tool to help with this process is handy.
The second major use-case of ICM is providing players with solid baseline strategies. We will usually know whether to play tighter or looser in a given spot, but its tough to know exactly how much tighter or looser we should play, and this is where the model can be useful.
Let’s take a look at a quick example of how establishing a baseline for a certain spot can be helpful.
Suppose these are the payouts with 3 left in a big online tournament:
1st – 15%
2nd – 11%
3rd – 8%
…and the stack sizes of the players remaining look like this:
Suppose further that the button folds and the small blind raises all-in, covering the big blind. With what range can the big blind call here?
First, take a moment to guess what the big blind’s calling range, then scroll down to see the results from the ICMIZER solver.
For this calculation, we will estimate the small blind shoves with this range:
This range includes 37% of hands. Note that this is a relatively tight shoving range. In theory (and in practice for good and aggressive players), the small blind can profitably shove as wide as any two cards here assuming the big blind has an appropriately tight calling range (which is another good baseline range to keep in mind).
Here is what the big blind’s expected value is with each hand against this 37% range:
Both AKs and 99 are losing calls, even though they have 67% and 65% equity against the shoving range, respectively. TT is a marginal call, winning just 0.02% of the prizepool on average.
Now, we have a baseline from which to adjust for this specific spot. We can use this knowledge to help estimate other all-in situations that are somewhat similar.
Getting acquainted with these baselines can be quite helpful. Seeing these spots over and over will lead to strong fundamentals when it comes to all-in situations on final tables.
Unfortunately, ICM gets A LOT wrong.
ICM is a very rigid model when the situations in which its used are very dynamic and situational. The model is very limited, which is no surprise considering it was originally created for horse racing.
The big issue is that ICM ignores obviously important strategic factors like:
We cannot rely on the model to adjust for these and other factors. Instead, we need to use logic in-game to determine the correct adjustments.
Another major limitation is that ICM assumes the tournament ends in one hand and therefore does not account for the future expected value of marginal and/or risky plays. It may make sense to go for a marginal/risky play if it substantially increases your likelihood of winning the tournament.
For example, if doubling up in a marginal spot would make you the chip leader, that marginal spot may be worth taking because you may be able to use your freshly grown stack to punish your shorter-stacked opponents.
In general, ICM tools are just too passive because of the future game limitations.
Lastly, ICM ignores the positions of the blinds, which is especially significant with a big blind ante in play (since you get a free orbit after posting the small blind). It may seem minor but the positions of the blinds can have an impact on strategies. Nick goes in-depth analyzing a hand in the ICM Unmasked course where two short stacks are about to be hit by the blinds and how this should impact every players’ strategy.
Nick sees ICM as a worse version of pure strategic thought. Rather than rigidly following ICM, we need to be much more dynamic with our strategy based on the specifics of the situation.
But even though ICM misses a ton of variables, you shouldn’t completely disregard the model. It is still beneficial for finding a solid baseline strategy in certain spots, but you need to be ready to adjust from that baseline. You don’t want to be an ICM robot.
If you want to keep reading, check out these 7 tips for tournaments.
Want to think about ICM like a Super High Roller champ? Get Nick’s ICM Unmasked course now!
This 7-hour course will change the way you look at tournaments for the better, and it includes Nick’s 4-hour review of a $2,100 tournament final table.
Leave a Comment