How To Get the Most out of Your Hand Reviews

My play seemed pretty straightforward, but I’m new to postflop ICM, so I wanted to be sure. Perhaps I should have 3-bet pre? Or check/raised the flop? Or bluffed the river? Or called the river? Or…

I marked the hand as one I wanted to review later, then let go of it in my mind so I could continue to focus on playing future hands as well as possible. I find that marking hands (or jotting down the details for future review, in live poker) is a valuable way to avoid ruminating on a mistake or potential mistake after the fact. I do want to learn from my mistakes, but the time to do that is not while I’m still playing! Marking the hand is my commitment to myself to review it later, which gives me permission to stop thinking about it at that moment and focus on the two cards currently in front of me.

Before you read any further, stop and review the hand for yourself. Is there anything you would have done differently? Now that I’ve told you the solver would have, can you guess what line it would have preferred and why? I encourage you to formulate your own hypothesis before you read on.

The solver shows A♥9♥ mixing in the occasional check. Note that it’s possible that the calculated EVs of mixed actions are not perfectly identical in a solution. This suggests it’s a close decision within the solver’s margin of error. In a vacuum, this is no big deal.

When comparing a presolved Chip EV sim to a custom ICM sim, however, we are actually changing two variables.

By running a cEV sim on the same preflop ranges as were used in the ICM sim, we can get a sense of how much the ICM considerations themselves contribute to the incentive to donk. It looks like the spot overall is bet at a frequency somewhere between the cEV sim and the ICM sim, with A♥9♥ specifically being a high-frequency but not pure bet.

We can also reverse this experiment by running an ICM sim with the Chip EV preflop ranges. That yields a higher donking frequency but is still well short of the 59% in the ICM (pre- and postflop) sim. A♥9♥ remains a mix:

This suggests that both factors contribute to BB’s incentive to donk-bet here. In other words, the ICM ranges make this turn more favorable to BB than it would be in Chip EV. The incentive of both players to conserve chips makes betting more appealing still.

How does betting help BB conserve chips? Let’s look more closely at how each of these factors shape BB’s strategy on this turn.

CO’s Preflop Range

CO’s ICM opening range is not wildly different from their Chip EV range, but one important detail stands out: the ICM range is missing some suited 7-x and some offsuit 9-x. That means they aren’t interacting as well with this turn card in the ICM sim, and in particular, they don’t ever turn the nuts with T9o. They do have T9s, but that accounts for fewer combos.

CO 25bb Opening Range

BB’s Flop Range

BB’s risk premium on the final table bubble is extreme. This discourages them from check/raising the flop for thin value with most of their top pair as they would in a Chip EV sim.

BB Flop Strategy vs CO’s Small C-Bet

CO’s c-betting range on the flop is not so constrained. ICM tends to be harsher on the out of position player, who will have less control over the size of the pot on future streets. More importantly, ICM is harsher on the player calling a bet than on the player betting.

Winning the pot without showdown is a fantastic way to avoid variance, which is exactly what ICM encourages you to do. The CO can win without showdown when they bet the flop, so they don’t need such high standards. BB cannot win (immediately) without showdown when they call, so they need a much stronger hand.

At the top end, BB calls more strong hands because their threshold for check/raising is higher. At the bottom end, they call fewer weak hands because their threshold for calling is higher. This all adds up to BB seeing the turn with a much stronger range in the ICM sim than in the Chip EV sim, while CO sees the turn with a slightly weaker one.

In the Chip EV sim, CO retains the range advantage at the top end. In the ICM sim, BB has more strong hands across the board, including at the top end.

ICM Discourages Slow-Playing

The other reason BB donk-bets so often on a favorable turn card in this ICM sim is that they don’t want to risk checking and losing the pot with a strong hand. Then, because they are betting so many strong hands, they can protect medium hands like A♥9♥ (which has 42% equity before betting, so it’s not exactly a bluff) by betting those as well.

BB never turns a set in the Chip EV sim, as they would have shoved 77 preflop. With T9 and 76, they mostly check the turn.

When BB turns a set or straight or two pair on the final table bubble, they mostly want to bet. Giving CO a free shot at rivering a straight or flush is an enormous risk at this stage of the tournament, whereas taking the pot down immediately is a pretty nice consolation prize if the bet does not generate further action.

Because BB rarely slow-plays this turn, they don’t have those strong hands in their checking range, which leaves their medium-strength hands more vulnerable to pressure from CO. A9s could still be the best hand, but it will be tough to get to showdown when it is (ask me how I know). By the same token, the presence of those monsters in BB’s betting range discourages CO from raising. By playing A9s the same way they would play their monsters, they not only bluff out stronger Ace-x but also make it more likely they win the pot when ahead.

Conclusion

I make lots of mistakes when I play poker. We all do. When I import a session into GTO Wizard and see those EVs pop up red, I don’t automatically review them. Often, I remember the hand and have a reaction like, “OK, yeah, I knew that was close.” I may even have deliberately deviated from the GTO strategy.

Over time, if I notice a pattern where I am routinely leaning the wrong way on a certain sort of close decision, that can be the impetus to study a spot more closely. But usually, those hands don’t warrant investigation, even if the pot was large.

This hand was a weighty one because it was on the final table bubble, but that wasn’t what made it interesting. It was interesting because:

The solver’s feedback prompted me to explore much bigger concepts.

In this case, the relationships between:

This was a case where I didn’t even consider the play that the solver picks the majority of the time. Nor did I expect donk-betting the turn to be such a significant part of BB’s strategy.

My frustration at the mistake was quickly overtaken by excitement. Postflop ICM has been my study priority lately anyway, and here was something juicy to sink my teeth into. Plus, I got good fodder for a blog article where I could share my findings with all of you!