Learn theory heuristics from one of the sickest hands played at Nosebleeds last year
The hand you’re about to see is INSANE.
It was a hand played at $200/$400 between 2 very strong regulars – Dmitry Grinenko and Taisto Janter.
More important than the hand itself though is what we can learn from it. And I think we can learn a lot.
Let’s dive into it.
The hand starts with Dmitry opening the Button for 2x sizing, and Taisto calls from the BB. The flop comes down J85 two tone, and Dmitry fires a 33% pot cbet.
There is already a lot to talk about here.
Dmitry used a 2x open size on the button, which is not a common sizing used at 6-max games with no ante. In fact, when the context is nosebleeds and tables with no ante, this betsizing is only utilized by regulars 8% of the time:
The reason for this is simply that the 2x sizing is not the highest EV sizing at equilibrium. When we compare the EVs of a 2x RFI sizing strategy to a 2.5x RFI sizing strategy, we see that the 2.5x sizing is about 8% more profitable:
Of course Dmitry knows this. Everyone that plays 200/400 online knows that a GTO solver prefers a 2.5x open from the button. That being said, some of them might decide to pick other strategic options for exploitative reasons. Most regulars know how to defend against 2.5x sizing, as that’s what they see the most and practice the most. But can they play appropriately against other bet sizings?
When we analyze the BB defense frequencies facing a BTN 2x RFI, and we compare nosebleed regulars against GTO, we do see a substantial leak.
While solver folds only 26.7% of the time and 3bets only 10.6%:
Nosebleed regs, when filtered for specifically 200/400 games (which are the same rake structure as the GTO print above), are folding 33% of the time and 3betting 14%:
This leak could in fact incentivize a BTN player to shift from the GTO 2.5x RFI sizing to the 2x sizing. At the same time, such a small sizing forces the BB to continue with a much wider range than usual, which is harder to play correctly and could lead to more mistakes postflop – something that would add even more EV to the 2x RFI option.
On the flop, we observe again a potential deviation from Dmitry. He fires 33% pot, a cbet sizing that is not used much at equilibrium on this flop, with a bigger 75% pot bet being much more used:
The small sizing is still used a little bit though, so there is not much we can infer about Dmitry’s strategy from this bet alone.
The hand continues, and now Taisto has to construct his check-call, check-raise and check-fold ranges.
On boards with lots of draws like this, construction of the check-raise range is usually very straightforward.
The first hands we are going to raise are the nutted ones – 2pair or better. In this case, J8, 88, 55 and 85 should all check-raise 100% of the time. Once the value region has been defined, we start adding bluffs.
The first region to draw bluffs from are the draws – and here we have plenty: flush draws of all sorts, gutshots (T7, QT, Q9, 96) and open-ended straight draws or double gutshots, in this case T9 and 97.
The next region of bluffs you should add into your range is the region of overcards with backdoors. These hands are necessary to include in your range because they will be your primary bluffs in runouts where your draws complete. Remember that you always need bluffs to balance your value region, so you need to prepare for the runouts where your main draws become nutted hands. In those circumstances you’ll need bluffs to balance those newly made value hands. And that’s where the backdoor hands will be helpful.
In this case, we are talking about hands like K9o with a club, KTo with a club, and all the Ax off with a club:
Lastly, the final region of bluffing hands you should include in your check-raise range are the bottom and middle pairs.
It might be weird to think of these hands as bluffs, and the truth is that they draw EV from all possiblities on the flop: folding out worse hands on the current street and getting protection; getting called by worse hands on the current street; and bluffing out better hands on next streets in favorable runouts.
I do however like to think of them as bluffs because I can anticipate that I should indeed blast off with them in some runouts in the XR-B-B line. The reason for this type of play is that these hands will become premium bluffs in these high aggression lines, as they will block sets and 2pairs for the IP player, making them better bluff candidates than missed draws from your range. In the Kh turn and 9d river runout for example, hands like 53 and 86o should blast off very often and go all in by the river:
Therefore, we see a flop construction that involves some of those bottom and second pairs in the check-raise range:
In the hand Taisto does go for the check-raise. Dmitry calls, and they both see a Js turn.
The turn pairing card is an interesting one. From experience, I observe a lot of regulars defaulting to a 30% pot sizing on these kinds of runouts, as if their range had become more condensed. However, it’s exactly the opposite that happens. The OOP range becomes more polarized on such cards.
As we discussed above, the OOP check-raise range is composed of essentially 2pairs/sets and bluffs. When the board pairs, the BB’s range is now mostly composed of fullhouses and bluffs; it becomes more polarized.
Of course, some 2 pairs will be counterfeit and become bluffcatchers, like 85 on this board. And the BB still has some trips combinations from having check-raised flop with some top pairs, which are not necessarily super nutted. But the bulk of the range will be nutted hands and weak hands:
Given how the paired card decreases the amount of possible combos of fullhouses available, the OOP’s range becomes heavy on trash hands like you can see in the print above. For this reason, the optimal play is to check a lot, and when betting, to bet big:
In the hand, Taisto checks and Dmitry places another 33% pot bet.
The small betsize makes a lot of sense for the IP player in these circumstances. As we discussed above, the OOP’s range is very heavy on trash hands. The IP player’s range however is very heavy on bluffcatchers – weak hands and good hands account for 75% of its range.
The matchup condensed and uncapped range x air heavy range leads to small bets being very much utilized. The reason for this is that middling strength hands benefit from making worse hands fold. This is The 3rd Reason for Betting, which I explained in a previous post here in the newsletter (if you haven’t read that post, don’t worry, you can access it later by going to my website homepage and scrolling down to find the list of all past issues).
Worse hands can sometimes improve and they can also sometimes bluff your middling hands. That’s the summary of why it’s actually a thing to try and make those hands fold. By making them fold you guarantee taking the full pot when you run against them, whereas by checking you not only allow them to catch their outs and beat you sometimes, but also you allow them to sometimes bluff you and take what should be yours (the pot).
To prevent that from happening, solver starts throwing these small bets to make air fold. It has to be small though, because your opponent’s range is still uncapped. So you pay a price to make those worse hands fold: you put some money into the pot in the occasions your opponent has a slowplayed nuts. If you had checked instead, you wouldn’t lose any money against their nuts. Therefore, to make this play work, you gotta risk only a little bit, just enough to make those air hands fold, while at the same time reducing the sacrifice of investing when your opponent actually holds the nuts.
This “trick” the IP player employs in these situations is so good that the OOP has to find a way to fight back. The OOP player needs to come up with a strategy that makes the middling hands from the IP’s range indifferent between trying to make OOP fold worse hands or just check for their showdown value.
That’s why you can see that all the one pair hands from IP in the above print are mixing bets and checks. OOP is playing in a way that guarantees they make the same amount of money betting small or checking.
The way OOP does this is by again developing a check-raise range. By bluff check-raising, the OOP player brings back exactly what the IP player doesn’t want: to have it’s middling hands bluffed out of the pot. The OOP player has to check-raise just enough to make betting small exactly the same EV as checking for the IP bluffcatchers. If OOP doesn’t do this sufficiently, IP can exploit him by always betting small with their middling hands.
In our hand, Taisto does exactly that. He check-raises once again, and Dmitry calls. They both see the 9h river.
The 9 is an amazing card for Taisto. QT, T7 and 76 all now made a straight. His full houses – 88, 55, J8 and J5 – barely lost any equity. And some of his turn 9x bluffs, like 97 and T9, now have a pair that can beat IP’s 8x combinations and some missed draw combinations.
The difference in range strength is so big that OOP has 59% equity on this runout, and captures over 60% of the pot:
OOP reaches this node with a significant diversity of decent and very strong hands:
- Some one pair combinations, like 97 and T9;
- Some Jx combinations, like QJ-AJ;
- Lots of straight combinations, like 76, T7 and QT;
- Some full house combinations, like 55, 88, J8 and J5.
Due to such diversity of hand classes, the GTO solution for the BB’s strategy includes utilizing several different bet sizes:
- The small bet sizes (20 to 33%) accomodate mostly the Jx combos;
- The medium bet sizes (50% to 75%) accomodate mostly the straight combos;
- And the all in size (125% pot) accomodates the full houses.
Looking at the above grid you can see that J8 and J5 almost never go all in, whereas 88 and 55 always go all in, even though they are worse absolute value hands than J8 and J5.
The reason for this are blockers: when you go all in with this line, the most likely hands that will call you are Jx trips combinations. With 88 and 55, you completely unblock that region, making it more likely that you can get called by worse when shoving. With J8 and J5 the opposite happens – you block your opponent’s calling range against shove, making smaller bet sizes a better option.
BB’s bluffs for the all in size include the flop bottom pairs, such as 53s and 52s, and also the counterfeit two pair 85s. Oddly, BB also bets all-in with some combos of 9x, such as 9d7d and Tc9d. More on that later.
Taisto goes all-in for 1.25x pot and Dmitry calls.
The bluffcatching strategy for the IP player is rather easy in this spot. Given the strength of it’s uncapped range and the fact that the OOP’s betting range is essentially straights and fullhouses or bluffs for this betsizing, many Jx combos are indifferent to calling or folding against this bet:
Some 8x combinations are also indifferent to calling or folding, such as A8o and 87s.
Taisto reveals 9c7s and Dmitry reveals…
A8o! Taisto finds an aggressive river play and manages to get called by worse. He scoops an $87,000 pot, leaving Dmitry with a sick hand history to tell his friends about.
This bet from Taisto may seem very odd at first glance. His combo is in fact a pure check in theory, but the difference in EV between checking and shoving is not so big:
97o combos without a club, such as 9d7s, should always jam:
How can that be the case?
These types of bets are called value-bluffs. They are a very rare occurrence at GTO. These bets are bets that can fold better hands and get called by worse hands at the same time.
In this case here, 97 is able to get hands like AA, KK and QQ to always fold, and sometimes it gets a fold from Jx hands like AJo, KJo, QJo, JTo and J7s. At the same time, it can get called sometimes by A8, K8 and 87s and win at showdown.
Value-bluffs appear in very specific range circumstances where the bluffcatching strategy for the opponent includes sometimes folding very high absolute value hands and calling low absolute value hands. This can happen when the bettor’s value range is extremely polarized, making hands like 3rd pair or trips pretty much equivalent in terms of strength. Then blockers may play a role where the hands in between – like 2nd pairs and overpairs – are not good bluffcatch candidates. If trips and 3rd pairs are being made indifferent at the same time, while 2nd pairs and overpairs are not good calls, then naturally a second pair hand can become a value-bluff – it will sometimes get trips to fold and sometimes get called by 3rd pair.
Here is another example from my Database of bot hand histories:
This bet with AJ on the river can get folds from better hands like QQ and Kx combos, while still having equity on showdown – it beats QJ that calls sometimes and splits with AJ that also calls sometimes:
SUMMARIZING WHAT WE LEARNED
With this hand alone, we were able to talk about:
- What’s the best theoretical RFI sizing from the BTN, and reasons to deviate exploitatively;
- Heuristics for constructing a flop bluff check-raise range (draws, backdoors with overcard and bottom/second pairs);
- Heuristic for playing the turn pairing card after flop check-raise – it usually polarizes the range more, while reducing amount of value combos. So proper strategy is to check high frequency and bet big when betting;
- The 3rd Reason for betting applied to the line facing-turn-check-after-calling-flop-check-raise. If the distribution of ranges is condensed and uncapped x air heavy, the optimal solution should include lots of blocking bets for the condensed range. This is because middling hands benefit from getting worse hands to fold;
- The way to fight back against the threat of many block bets from middling hands is by developing bluff check-raises. Not doing so can be exploited by the opponent in the form of always betting the middling hands to fold out the air;
- Why many bet sizings are often necessary at equilibrium: if you have many different regions of potential value betting hands, and they all have substantial size (number of combos), you likely need many different bet sizing options to maximize your EV;
- Blockers play a role in what’s the best bet sizing for your nutted hand. Unblocking calling range leads to a size up, while blocking calling range leads to a size down;
- When your range is sufficiently strong and protected, you can afford to fold very high absolute value hands, like trips against a near pot sized bet;
- Sometimes value-bluffs appear in GTO. These are bets that get called by worse sometimes while also folding better hands sometimes. They happen in rare circumstances when the bettor’s range is extremely polarized and blockers play a significant role.
If you would like to see more breakdowns like this in the newsletter, please respond to this email so that I can know that you liked it!
| For more high stakes breakdowns, check out this video where I reviewed the Top 3 Sickest Nosebleed hands from February/2024
Thanks for reading. See you next week.
Until then – keep it simple.
Saulo