POTD #108 In The Triton Million I Roll The Dice vs Mario Mosböck

A very deep stacked hand in a $500k Tournament

NB: I have started a subscriber chat and have launched it with an AMA for all (not just paid) subscribers. If you are interested, you can join the chat here

Something I’ve alluded to in many posts, but haven’t directly talked about, is randomizing. It’s something all poker players do intuitively, and many top players, myself included, do systematically.¹ The idea is that in NLHE, you will often have a hand that wants to play a mixed strategy, and to make sure your game isn’t too predictable or exploitable, you will have a mechanism that gives you a random number and you will let that number dictate your play. For most players, a low number indicates a passive play and a high number indicates an aggressive play. Some players randomize using a system that includes every number from 0-100, but my system uses intervals of 5 from 0-100. I can’t pretend that my strategies are so specific that I know whether I should be raising 40% of the time or 44% of the time. I will not tell you what my exact system is because I am always paranoid of someone discovering it, seeing I rolled a 100, and then calling me light, but it’s not all that complicated, and I have a variety of different systems that can change depending on the tournament or my mood.

Some reading this might say, to use a Will Jaffe coinage, why would I need to randomize when I’m playing “Billy from Poughkeepsie”? I think there are many spots versus weaker players where you do not need to systematically randomize. If a guy who has three-bet once all day looks really excited and three-bets you, you do not need to randomize four-bet bluffs with A5s. However, you also want to keep your overall game plan reasonable. If a weaker player raises the LJ, are you going to three bet every suited ace 100% of the time from the HJ? If so, you need to recognize that the CO/BU/SB/BB and LJ all might pick up on what is happening and play back at you. Once you decide something like “well sometimes I’ll three-bet ATs and sometimes I’ll call it,” using a randomizer as a tiebreaker for that individual decision has some merit.

If you are playing versus expert players who are actively trying to exploit your tendencies, I think randomizing becomes even more important. One’s natural poker gravity will usually lead them towards certain plays; when playing vs. experts, you need to self-correct to prevent picking the same sorts of plays too often or too rarely. If you bluff every 5% bluff combo on the river, each bluff technically is breaking even according to the solver, but your overall strategy is so out of whack that an observant player will start calling every bluff-catcher and your bluffs will start losing a ton of EV. I would not recommend every reader of this blog randomize in every hand, just as I would not recommend every reader of this blog consider hero-calling the river vs. every single opponent they ever play with, but randomizing is a powerful tool that can help you in tricky situations vs. top competition, and learning to implement it in your game will have benefits.

Bahamas Triton Million (Actually $500k)
1M Starting Stack. Level 1 (2k/3k/3k) (SB/BB/BBA)

HH: Triton App

Mario Mosböck (1M) raises the LJ to 8k, I (974k) defend K♠️8♠️.

Flop (21k) 9♠️7♠️6♣️: I check, Mario bets 14k, I call.

Turn (49k) K♥️: I check, Mario bets 55k, I call.

River (159k) 3♥️: I check, Mario bets 225k, I fold.

What I was Thinking

In my notes for this hand I wrote my rolls down, and I rolled a 40 on the flop, a 5 on the turn and a 20 on the river. I did not write down my preflop roll, because I thought I had a pure call, but it turns out I get some low-frequency three-bets. This tournament is an invitational that began with split pro and VIP fields: pros vs. pros, VIPs vs. VIPs. At the end of day 1 the fields combined, so my overall gameplan involved playing less aggressively to make it more likely I made it to day 2, so pure call preflop seems fine. I know that on the flop I thought I mixed raises, but I rolled a 40 and thought a 40 was right on the cusp of being a raise, but decided that since I was playing more passively in an invitational that I should call. I also did not think Mario was supposed to play big bet on the flop very often with range, and I thought he was marginally more likely to be greedy with a top of range hand than setting up a huge bluff. On the turn, I thought I had a hand that would occasionally raise and could get better one pair hands to fold, but I couldn’t raise it on such a low roll. I was also concerned that Mario’s turn bet was not solver-approved and considered he might be being greedy— it’s a normal bias for people to find outlier sizes more often with the nuts than with bluffs. On the river, I felt I had an adequate bluff catcher, but my low roll, combined with the situation discouraging aggressive play and my reads on the flop and the turn, pushed me into a fold.

What I Got Wrong

I was wrong about Mario’s flop size. He doesn’t always c-bet this flop, but when he does, 2/3rds pot is an appropriate size with range. I should raise the flop on a 40 and the only 8x+FD that would not raise on that roll are special combos like A8 (NFD), T8 (straight), and 86 (pair). However, calling the flop is fine if I want to decrease my aggression a little because it’s an invitational or if I think Mario is a little overweighted to value.

I was right about Mario’s turn size. B120 is never used with range, and his most common size should be around 2/3rds pot again. When he bets 2/3rds pot again, Ks8s is a pure raise and raising gets KQ and KJ to pure fold, while also generating some AK folds and occasionally getting called by hands I dominate like 88. If I only give Mario one turn size, b112, he still bets the turn frequently, almost half the time with range. I can raise 80% pot and now I get AK and AA to mostly fold, or I can pick a giant raise size like 150% pot and get sets and two pair to start mixing folds. In practice, I don’t think I am ever getting sets or kings up to fold, but I think if the smaller raise size would get stronger one-pair hands to fold while also getting called by draws I have dominated and I think that is what I should have done. A problem with randomizing is you often lose the forest for the trees; it’s possible if someone asked me what the best play on the turn is, I could have said raising as a merge, but once I rolled a low number I needed to be very confident to override my low roll.

On the river, I didn’t consider leading with my hand or range, but it’s frequently used, as I have a lot of strong top pair and some draws that are total air with no showdown. I can lead a hand like mine on the river and hope to get called by a hand like T9 while slowing down Mario’s bluffs. Given that I felt his sizing on the flop and turn had polarized him quite a bit, I don’t think leading would make sense with my reads, because he probably has fewer combos of hands like T9 or TT than he’s supposed to. Once I check, the smallest river size he uses is 140%, and he can bet as large as 565% all-in. All the K8 hands are breaking even or losing a very small amount to call on the river; having the Ks is slightly bad because it does not block two pair, and that lowers the EV of calling by 1.5bbs, which sounds like a lot, but is not that much when your opponent is betting 75bbs into a 53bb pot. Given my general reads about his flop and turn betting strategy and that my hand is a pure fold on the river in the sim I ran, I am happy with my fold.

Types of Errors

Failure to merge

Grade

I paired this hand with POTD #62, where Andras Nemeth found a really tricky and nice merge three-handed in a SCOOP PKO. In today’s hand I considered merging, but ultimately chickened out. Was this a reasonable pivot because Mario’s range was too strong? Given how he played the rest of the day, I don’t think so. I still think this was a reasonable deviation in an invitational, even if it was a single re-entry tournament, and while Mario ran some big low-frequency bluffs throughout the day, this turn check-raise really only works if he’s consistently playing top pair like this. If he’s overplaying two pair and sets on the turn, just calling is probably the right deviation. I don’t think I made a clear mistake in this hand, but I feel I let randomizing cloud my thought process throughout, so I will give myself a

B-

1. It’s also become a more common semi-systematic play where people flip a coin to decide what to do. Flipping a coin to determine your actions in a spot where you believe you have a mixed strategy? That’s using a randomizer.