POTD #178 Winning Wednesday: Maria Konnikova's NAPT Punt

A hand from the $10k High Roller before her Second Chance victory

This week we are doing something a little different on POTD. For POTD #47, I wrote about a hand Maria Konnikova played in EPT Monte Carlo vs. Dimitar Danchev. This week, we will be looking at hands Maria and her Risky Business co-host Nate Silver played in NAPT Las Vegas at Resorts World, and I will be declaring one of their punts the biggest of the two.

Fortunately for readers of POTD, they made life difficult for me, but not by punting in such spectacular fashion I couldn’t determine who was the bigger punter. I was not torn between wondering what’s worse, jamming 72o UTG or three-bet shoving 32o because they “had a read.” Instead, they made life difficult for me by playing a variety of interesting poker hands where I may have punted in a similar or different fashion if I were dealt the same cards they were. Today is Winning Wednesday, and while Maria did not cash Event #42 the NAPT High Roller, the tournament we will be writing about today and tomorrow, she did end up winning Event #48, the Second Chance NAPT High Roller, which put her over $1,000,000 in lifetime live cashes. Congratulations! Her win means she deserves the honour of having her hand written about on Winning Wednesday, while Nate will have to settle for the ignominy of having a hand written about tomorrow, a day no one in the USA notices or cares about.

NAPT High Roller $10,300 Event #42
(1k/1.5k/1.5k) (SB/BB/BBA). 50k Starting Stack. Registration is Open.

Maria (65k) makes 3k with A♣️5♣️ UTG8, Thomas Eychenne (120k) calls next to act, it folds to Barry Woods (35k) in the BB who calls.

Flop (11.5k) Q♣️7♣️2♦️: Maria checks, Thomas bets 3k, Barry calls, Maria calls.
Turn (20.5k) 5♠️: Barry leads 4k, Maria makes it 24k, Thomas shoves, Barry calls, Maria calls and loses to Barry’s 75 and Thomas’s AQ.

It is Thanksgiving week and I am in the giving mood so today’s and tomorrow’s post are free for all. If you’d like to subscribe to POTD you can get 40% using https://www.puntoftheday.com/BlackFriday

What Maria Was Thinking

Maria was kind enough to share her thoughts with me and they are quoted below. They have been lightly edited for formatting and clarity. [Her remarks have not been edited because there were no mistakes. Maria gets an A+ for her grammatical work. -ed.]

Preflop, I’m raising that every time. Flop, I usually check to IP here, since he does have a lot of Qx. I didn’t think there was a particular reason to raise after Thomas bet and Barry called (though maybe you disagree!). Where I think I started punting is the turn. Barry’s lead just really threw me off. It seemed like such a random place to do that - and I decided that with a pair and nut flush draw, I’d raise and get Thomas to fold out some Qx that beat me, if not on turn then on river. Thomas is super aggressive but I’m still surprised he shoved the AQo there. Curious if he’s supposed to. Anyway, after he shoves and Barry shoves, I’m obviously beat by at least one of them (I actually thought Thomas was likely to be on a draw when he jammed) and just need to figure out if I want to try to hit my outs. With over 150k in the middle and me with less than 35k behind, I decided I had to call. I put myself in that nasty spot all on my own and I think I should have found a way to avoid it.

What I Thought (no cheating outside of knowing the results of the hand)

One of the privileges of doing POTD every day has been that I’ve been running a lot of sims on hands that are odd. I tend to recommend that students get very good at playing single-raised pots, preflop raiser vs. BB. It’s a good way to learn solver mechanics, and you are studying the most common types of pot you’ll play and the types of pots with the largest database of solutions. All of this windup is to say, because of POTD, I’ve done a lot of work in multiway pots, and I believe that the preflop raiser plays very heavy check three ways on queen-, but not jack- or king-, high boards EP vs. EP, and that once the flop goes bet/call, Maria will mostly play raise or fold with her entire continuing range, but specifically she will almost always raise nut flush draws.

No amount of studying in multiway pots could prepare me for what to do when the BB leads the turn three ways on a card that appears to be a total blank. My first instinct was this was a good raise from Maria; the check-call/lead on a blank does not appear particularly strong. Maria has a very strong draw and she might be able to get either player in the hand to fold a better hand. The problem is, she’s not really representing anything. She’d always raise the flop with an overpair, set, or strong top pair. She would never call the flop with a hand like 55, so her hand really looks exactly like what it is. It’s always possible to baffle people at odd nodes and force some huge folds, but I would prefer calling.

What We Got Wrong

Preflop is as standard a raise as you’ll get, and Maria and I both correctly identified the flop as being a high-frequency check. However, once Maria checks, Thomas bets and Barry calls, the solver agrees with me that Maria should raise. To this action, she folds 57%, raises 36%, and only calls 7%. Her exact combo mixes calls, but nut flush draws in general raise a lot more often than other flush draws in her range, even if A2 and A3 are the only combos where there is a significant EV difference between calling and raising. I still favour raising for a lot of reasons. You can get many better hands to fold right away: Thomas will fold hands like JJ to a check-raise right away, and Barry can fold top pair if Thomas calls Maria’s check-raise. Additionally, if we overcall, I think our range is face-up— hands like 7x, middle pocket pairs with a club, or flush draws— so when a club hits it might be hard for us to get paid. Folding out better hands right away and making the stack-to-pot smaller so we can win a really big pot if we hit our draw seems ideal, and playing strong draws the same way as you’d play strong made hands seems like the best strategy.

The turn surprised me quite a bit. Barry leads the turn almost 40% of the time, including 100% of the time with 75. The logic of his turn lead goes like this: Thomas can stab the flop with all sorts of garbage, hands like JhTh. Barry never raises the flop with Maria’s uncapped UTG range still to act behind, which means Barry has a lot of top pair, two pair, and strong draws. Once Maria overcalls the flop, she mostly has middle pair and draws. Barry has an uncapped range acting first and facing a range that is very capped and a range that has a lot of total air; it follows that he would lead often. Once Barry leads on the turn, it’s confusing, and when I am confused and have a big draw I tend to raise (eg. POTD #101). But the unflappable solver recognizes our range for what it is— a bunch of middle-of-range junk— and it never raises anything when a (mostly) uncapped range leads into us and there is a (mostly) uncapped range behind.

Once Maria raises, my initial thought was it was too large, but I actually think it’s a good size. If she were pure bluffing here with a hand like A7, I think she’d want to make it smaller so it’s cheaper when Thomas goes all-in or so that she could maybe raise-fold to Barry. In this hand, the worst case scenario would be if she raised to, say, 12k, got shoved on by Thomas, and then needed to fold. Calling is still the best play on the turn, but if you have to raise, I think you want to make sure you’re never folding such a monster hand. If she had a hand like Ac6c, she might even be able to fold to this action, but I think there’s too good a chance an ace or five scoop the pot for her, so she needs to call off.

What Sam Thinks and Grade

Maria made two mistakes in this hand: not raising the flop and raising the turn. Playing heavy raise on the flop with range works out well vs. humans. It’s not just a tricky solver play; playing aggressively with strong hands in multiway pots is usually the best way to extract the most value out of them. That being said, even the combinations that always raise the flop here don’t make a ton of EV compared to calling. The real punt would be doing something crazy like folding Ac5c on the flop, but just calling makes life hard on your range on future streets and also loses EV vs. the solver and most human players.

The turn is an example of getting lost and not quite recognizing your relative position in the hand. Let’s briefly look down another turn node: the one where Barry bets and Maria calls. Do you know what Thomas does with Ac5c? He raises it every time and makes almost 3bbs compared to just calling with it. Ac5c is a very good hand to raise on the turn for all the reasons Maria thought. The problem is, she can’t credibly represent any strong hands herself, which means she can’t get Thomas or Barry to make any big folds, which means she’s usually just putting in a lot of money with 30% equity or less, which is not ideal.

C