Week in Review #33 November 3rd - November 9th
Learning the balance between having an open mind and just being a contrarian
Last week on POTD was Halloween week. I spent all week writing about the horrors of getting coolered in poker tournaments, I put my final touches on a Week in Review post about the rigged poker game scandal involving NBA players, and prepared myself for a different type of horror: The Toronto Blue Jays losing two very winnable games that could have won them the World Series. This will be my last baseball writing for a while— if POTD subscribers want to read 300 words about Nick Sandlin, sound off in the comments— but before I wrap up baseball blogging, I wanted to write about a trend I’ve noticed in poker, sports, and society at large, which is an inability to calibrate when one should deviate from some rather simple technocratic heuristics.
As a teenager, my online time was spent playing online poker and reading Two Plus Two. It was easy to make money at poker, in part because many of your opponents did not understand simple math. Someone shoves 10bbs from the SB, you should call K5o; if your opponent ever folds K5 or hands like it because they “don’t want to gamble” or because “getting all-in preflop isn’t real poker,” you could get away with shoving almost any two cards. There was similar easy money to be found in sports; walking is good, bunting is bad, three points are worth more than two, don’t punt. These beliefs trickled through nerdy internet subcultures, before much of that nerdy internet hive mind took over the world. They were hired by sports teams, correctly predicted presidential elections, or in the case of poker, ended up becoming the best players in the world.
This trend has perpetuated to the point where it has robbed some of these sports of some stylistic variety, and there has been an increased homogeneity in many arenas that teeters on boredom. Would you rather watch a basketball team do everything in their power to generate a mediocre three point attempt or watch players with amazing handles show off their skills trying to dribble through an entire team? There will always be edge-seekers trying to buck conventional wisdom, and a lot of contemporary conventional wisdom was formed in the early 2000s. We at POTD always support trying to be one step ahead of the curve and adapting to new meta game conditions and also tend to avoid grand sweeping statements about the world as a whole, but I fear what we are seeing is a trend of people misusing data to poke holes in the current consensus as an excuse to revert to older, dumber strategies.
A good example of this, which I discussed on my episode of the GTO Lab Podcast, is the “Hot Hand Fallacy.” People who played basketball (or any sport) swore that there were times when they were “locked in” and they would be more likely to hit a shot than their career averages would say. Fans of basketball, claimed that they too could tell when players were in the zone. In 1985, there was a paper written debunking the concept of the Hot Hand Fallacy, and it was affirmed by a 2003 study that also claimed that streak shooting was almost entirely explained by dumb luck. In 2018, there was a study that debunked the 1985 and 2003 studies by showing, amongst other things, that the math of counting streaks is a lot more complicated than you might think (read a thorough summary of this here).
At one point in time, I may have believed the hot hand effect was a “fallacy,” and I have been convinced it is not, but the question that matters when you are making a decision in game is not “does it exist?” but “how big is the effect?” I fear we are in a moment where there is a lot of smart-alecky, knee-jerk contrarianism— and trust me, I would know, I am a smart-alecky contrarian— and confidently asserting the “hot hand effect” is real does not mean you should, say, let a hot Brandin Podziemski shoot the game winning shot over a cold Steph Curry. This brings us to the bane of my existence the past couple months; John Schneider.1 [The Blue Jays’ manager-- I don’t have reason to believe Sam has a beef with the Seattle Seahawks’ general manager, nor the actor/singer who played Bo Duke in The Dukes of Hazzard. -ed.]
Modern baseball is inundated with all sorts of new tracking data. They track how fast batters swing and what angle their bat is at, how often the ball spins when pitchers throw it, they measure velocity down to a decimal point— does it really matter if a pitcher is throwing 97.4 or 97.7 MPH? When Schneider was hired as a manager, he was lauded for his ability to combine quantitative and qualitative elements in his decision making by The Athletic:
In the dugout, Schneider introduced a more aggressive managing style, especially on the basepaths and in the ways he deployed his relievers. And, in Schneider, the Blue Jays have a manager who combines data with a feel for the game.
“One of the many attractive things about John is how prepared he is that allows him to be agile in a game. There’s so much work done prior to think about all the different hypotheticals. You can’t possibly think about every one of them. We certainly try to when we have the opportunity,” Atkins said. “But agility is huge and being able to rely on experiences and ultimately trust your process to make decisions in the moment has to be there and it was evident to us that he was prepared to have that confidence to be agile.”
I believe he has reams of data and plans out every game for all sorts of hypotheticals, but when he violates the basic tenets of early baseball analysis— don’t bunt, don’t intentionally walk players— more than anyone in baseball, I fear what we are seeing from him is using all the information at his disposal to overcorrect when the right answer is the conventional one. Unlike using players on short rest, strategies like bunting or intentionally walking are relatively costless, in that they can be implemented in the preseason as easily as they could in game 7 of the World Series. So when a manager does it more often in the playoffs than in the regular season, it usually is not a sign of high-level thinking at the highest stakes, but panic. The stakes are so high that he can’t sit there and do nothing; he needs to tinker, he needs to be active.
I often see similar risk-averse behaviour deep in poker tournaments. Sometimes you have 88 and the other player has KQs, and you get all in and there is nothing you can do about it outside of hoping you win the flip. No amount of preparation or extra information could change that. You could run an FGS-1000 sim on a quantum computer until it converged to a perfect solution and it would still say “you need to get all-in here.” Using big data or counter-intuitive logic to justify a bad play is the product of scared decision making; it does not represent the next frontier in expert decision-making.
If you were able to play exactly like a solver in every single situation, you would be guaranteed to win in the long run, but you would not necessarily be the biggest winner in a given game. When solvers were first introduced, they were presented in apocalyptic terms for the future of poker: “These bots can outplay any human and are unbeatable.” Recently, I’ve been seeing a lot of poker content that is moving in the opposite direction— yes, a solver would beat you, but you are not playing a solver; you are playing a flesh-and-blood human with repeatable, predictable leaks that you can exploit in a way that a solver cannot.
The best way to make money at poker, especially at lower stakes, is by exploiting your competition and not playing perfect solver poker, but this is not an emperor-has-no-clothes moment for solvers. There will be spots when you don’t know what your opponent’s range is and you should revert to solver baselines. There will be spots where you might notice your opponent is c-betting too often, so you start check-raising a lot, and then they start adjusting by three-betting more flops or c-betting less frequently. Not every hand vs. every human will have an easy max-exploit strategy that you can uniformly and repeatedly execute. You will occasionally want to play an alternative strategy, but those hands should be the exception, not the norm.
I think it’s interesting and beneficial that the “Hot Hand Fallacy” was debunked, but the original 1985 essay was also a necessary corrective to conventional wisdom that greatly overstated the effect of having a Hot Hand. Punting (in football, not poker), bunting, intentional walks, and mid-range jumpshots have all become less common; this represents progress. It’s possible there has been an overcorrection and we are due for a slight comeback, but it should be viewed as such, as a slight correction in an ever-changing meta-game, not an opportunity to proudly return to more ignorant times.
If you’d like more writing like this from me, or just good old-fashioned hand history analysis, please consider becoming a paid or unpaid subscriber. I know a guy playing $100,000 poker tournaments asking for $10/month might seem a little greedy, but I am writing around 10,000 words a week, which is around the number of times I’ve replayed the 9th inning of game 7 in my head. All the subscription tiers are explained here, and you can subscribe or upgrade your subscription by clicking the button below.
Additional Sims For Premium Subscribers
Premium subscribers get the raw files of sims I used to write my POTDs, sims that are more accurate and appropriate than equivalent sims in the big public libraries, videos of me walking through the sims, and a text summary of how I ran the sims. This week I uploaded:
A PIO sim using alternative preflop ranges for POTD #162
Three different PIO sims taking us down different branches for POTD #163
Some TBD sims for the very tricky hand in POTD #164
A PIO ICM sim for POTD #165
Additional Analysis for Premium Subscribers
Everyday Premium Subscribers get an extra bit of analysis not included on Substack. Today I’ll share #onemorething from POTD #162, where I wrote about bet sizing tells vs. unknowns.
I remember playing an online high roller years ago and OTB_RedBaron overbet the turn in a spot where I didn’t think one was supposed to overbet the turn and it blew my mind. I could not fathom what range to put him on or what hands he was representing.
Now that most of my poker play is in high rollers, I have an opposite problem where I don’t necessarily realize what a bet size from an average main event regular is representing. In today’s hand, 55k into 85k represented an unclear size to me; it seemed a little small for a T. I think there is often a meta in high rollers where there isn’t quite enough slowplaying and people usually bet too often and too big with good hands. To me, 55 into 85 represented better than a T, but a thought did cross my mind: What if he is betting hands I chop with like AJ or A7 or hands that I block like AK or AQ with this size on the river? Then folding top pair would probably be a disaster, so I need to call. Upon further reflection, I think it’s more likely that Jason wasn’t going to overbet the river with anything, and this was probably just his standard value-betting size that would include straights, trips and even full houses and would rarely include aces up, so I maybe could have called because the small size gave me good pot odds, but not because I chopped with value bets.
Housekeeping
The POTD Discord is now public. If you’d like to join it, feel free to do so here.
If you’d like to sign up for Octopi, Run It Once, or GTO Lab you can get a discount using the following codes.
Run it Once use code: POTD for 10% off
Octopi Poker use code: PUNT for 50% off 1st month for monthly subs and 2 free months for annual subs
GTOLab use code: POTD for $25 off any product. It can be used multiple times
Media
No media appearances from me and not much media watching from me.
As always you can reach me
I took a psych class at Cornell from the guy who wrote the 1985 paper on the Hot Hand Fallacy. Separately, he gave a lecture on risk aversion and asked the class if anyone wanted to flip him for $20. I got him for the $20