This Radiolab segment on poker is worth checking out. As usual, they get a few details wrong, but present the big ideas in a compelling way. If you’re not much into poker this is a good introduction to thinking in terms of probabilities.
grindrewind
thoughts on poker
Old Men Behaving Badly
Claude*, a regular in the game, a sixty-ish gentleman who wears a branded golfing cap, a polo shirt, and several layers of cardigans (all collars popped), was banned from the casino for a day for “throwing cards and chips at the dealer.”
Wouldn’t it be nice to just let it all go?
My late grandmother played the Italian card game Scopa very well, but when she lost she threw the cards and accused the rest of us of cheating, a habit I acquired for a time. I still respect it: if you’re not freaking out, you’re not trying. But these days I keep an evener keel. Barry Greenstein says the rules are different for winning players, and they are. Noblesse oblige.
And yet playing with Claude always threatens to disrupt my equilibrium. Watching him squirm elicits schadenfreude so strong that I constantly catch myself contemplating plays not because they’re +EV, but because of how mad they would make Claude. And I’m not the only one.
I wasn’t around to witness this hand. I heard about it from Chad and was instantly humbled, because, as much as I yearn to mess with Claude, I would never in a million years dream of going half as far as Peter did in this hand.
It happened in a 5/10 PLO game with a $20 straddle. A few players limped and Chad (who told me about the hand) called in the cutoff with 9987 with two clubs. Claude raised to 60 on the button, Peter called in the big blind, the limpers called, and Chad called. The flop was Ac Ks 9s with $300 in there from the pre-flop action.
Everyone checked to Claude, who bet $200. Peter called, the limpers folded, and Chad called. Claude is a very passive player, so when he bet the flop into four players, Chad wasn’t feeling great about his bottom set, but didn’t want to fold just yet. After calling the $200, he only had $1200 left, whereas — and this would become very important — Claude had over $5000 and Peter covered.
The turn was the 6c, giving Chad a backdoor flush draw and open-ended straight draw. It checked to Claude again and he potted for $900 and Peter called. Once Claude potted the turn, Chad was pretty sure his set was no good, but with the flush draw and open-ender, his hand was too strong to fold. He called the $900. With only $300 left behind and $3600 in the pot, he was effectively committed.
The river was the Qh. Peter led pot for $3600. Chad put in his $300, more or less resigned to losing his stack, but unable to fold for the last few hundred. Then the action was to Claude. He has a routine he goes through when the tide of a hand turns against him. He breathes heavily, mutters to himself, holds his cards up to his face, and flicks them. After several minutes of that he folded his hand face-up: AKK5 with clubs, for a set of kings with the nut club draw. Peter showed QsTdThTc for a pair of queens. Chad’s set of nines won the $4200 pot.
In case you’re unfamiliar with Omaha, let’s break down Peter’s play in this hand. First of all, he called a pre-flop raise with trips. Trips are the absolute worst hands in Omaha. Since you have to play exactly two cards from your hand, you can’t play pocket trips, and having the third card of the rank in your hand decreases the chance that you make trips on the board, as well as making it impossible to make two pair. Even people who play every hand don’t play trips.
Then Peter flopped a bare gutter — basically nothing — in a five-way pot, but he still didn’t fold. He probably called the flop because he had the Qs, meaning that if a third spade came he would have a blocker to the nuts, which he could use to bluff. Calling for a draw is often a bad idea: you have to consider the odds, how many cards you’re likely to see, whether you’ll get paid off, and so on. Drawing to a bluff is just bonkers. But if the flop call was bad, the turn call was much worse.
When the river came a heart, Peter missed his spade “draw.” But he seized upon the opportunity to use his three blockers to jack-ten to represent the nut straight. The only problem was, by this point Chad was already committed to the hand. When one player is all-in, it’s known as a protected pot. There’s no reason to bluff at a protected pot because you still have to win at showdown against the guy who’s already all-in. Thus, Peter’s bluff could not possibly make him money — all it could do was screw Claude. And this is what made it so credible from Claude’s perspective. Since there was no rational way Peter could be bluffing, Claude laid down a huge hand.
I’ve never met Peter, so I can’t say I know for sure why he played the hand the way he did; but based on the action, the spitefulness of his line is staggering. He hung around till the river and invested $1200 for the opportunity to run a bluff that couldn’t possibly succeed. It would be like running an Iron Man for the sole purpose of tripping someone at the finish line. This hand is the Bhagavad Gita of trolling.
After years of learning to be robotic, to focus on the odds and the strategy, and nothing else, of murmuring graciously, “Nice hand,” when someone beats you, no matter how improbably, there is a certain release in watching someone emote freely, scream and curse, throw chips and cards. So, hats off to you, Peter, you bastard. I wish I could’ve been there to see it.
*Names changed to protect the spiteful and the wrathful.
Clash of the Titans
I got the first message, looked over, and saw Dom sitting to Will’s left at 5/10 NL (I was playing PLO). Naturally I hoped my two friends would get into some big pots that I could later dissect and they didn’t disappoint. They soon played an interesting pot that I heard about from both players individually.
Two passive players limped in early position, Will raised on the button to 50, and Dom three-bet to 185 from the blinds. The two limpers folded and Will called. The flop was Q42 rainbow, Dom bet 200, Will raised to 450, Dom went all-in for 1250 total, and Will folded.
I heard about the hand first from Dom, who had AK. His three-bet pre-flop with AK is completely standard, of course. On this dry flop he’s going to be continuation betting most of the time. Betting around half pot is also standard here: he can still easily get stacks in by the river and there’s no need to risk more on this texture. Naturally he would use the same sizing whether he was value betting or bluffing. When Will raised, he thought he was representing a very narrow value range of QQ, 44, 22, and maybe slow-played AA or KK. He didn’t expect Will to raise with KQ or AQ, and in fact didn’t expect him to raise with sets either, since most people would slow-play them on such a dry flop. Having ruled out most value hands, he concluded the raise was most likely a bluff and went for the re-bluff.
Since Will raise-folded the flop, his hand ended up looking like a bluff, but I had an inkling that he might have had a marginal made hand like JJ, TT, or even a weak Q. As it turned out he had QT suited. I would over-limp this hands sometimes, but raising passive players in position with a fairly strong, playable hand certainly can’t be bad and it’s what I would do most of the time. Will probably expected Dom to be three-betting wide, since Dom usually gets more than his share of three-bets in and I assume this session was no different. But I actually don’t think Dom would be getting out of line here. He has a lot of history with the players who limped in early position (although there’s no way for Will to know this) and they would be cold-calling him here a decent amount; plus, he probably saw Will as one of the tougher players at the table and wouldn’t go out of his way to get involved in a big pot out of position against him. Overall, it’s just not a great spot and I don’t think Dom would be very likely to try a raise without a strong hand. Given all that, Will could have folded QTs here. It’s a nice-looking hand, but it’s dominated against a tight three-betting range of, say, TT+ and AQ+, and stacks aren’t deep enough to maneuver too much post-flop. But this reasoning depends on some reads he couldn’t have been aware of, so I can’t really fault him for defending QTs in position against an aggressive opponent.
The flop showed why defending a hand like this can be dangerous: Will flopped about as well as he could hope to, top pair, but still couldn’t be thrilled about his hand. He raised in an effort to slow Dom out and perhaps gain some information, but once Dom shoved Will was convinced he had at least an over-pair and mucked his hand.
As I’ve discussed before, raising to “find out where you’re at” is dangerous against tough opponents. The problem is that if they don’t react honestly you’ve built a big pot without gaining any information. In Dom’s shoes some players would just give up their bluffs without thinking too much; others would be suspicious, but wouldn’t pull the trigger on a re-bluff for their whole stack. Will was unlucky to run this play against the worst possible opponent, in that Dom would perceive the raise as bogus and do something about it. In fact, I think Dom would be so convinced that Will’s raise was a bluff that he wouldn’t shove if he actually had a value hand; he’d call to allow Will to continue bluffing. So once he shoves the flop, he’s almost certainly bluffing.
Arguably Will’s only mistake in the hand was folding to the shove, but unfortunately, it was a big mistake. Many players just ask themselves, “Am I beat?” but it’s far more valuable to think in terms of equity. Once Dom shoves, there’s a total of 1900 in the pot. Will has to put in 800 more to call. His break-even equity is 800 / (800 + 1900) which is 29.6%. Against an over-pair, he still has five outs, which gives him about 20% equity. He’s not so far from being priced in against Dom’s value hands, which means Dom doesn’t need to be bluffing all that often to make this into a call. Based on the above reasoning, Dom is actually bluffing quite often here, making it a clear call.
Which means that Will’s raise would have been brilliant had he been planning to call it off. This is known as raising to induce. A lot of good players have no raising range on a flop this dry, but I think that’s a mistake. Dry flops are unlikely to improve either player, which means it’s pretty likely no one has anything. The battle of nothing versus nothing is the most important battle of all in no-limit hold’em. Taking raising out of your strategy makes it a lot harder to win this battle. Raising with a balanced range of bluffs and value hands is a powerful strategy on dry boards as well as wet ones. I think a good strategy on this flop would be to raise and call it off with top pair-bad kicker, and call down with top pair-good kicker (KQ/AQ). This has a few advantages. If your kicker is low, you’ll have five outs against an over-pair, giving you about 20% equity, whereas if you got it in with KQ v. KK, for example, you’d only have two outs for about 8% equity. Additionally, a K or A on the turn is more likely to get bluffed than a low card, so your bluff-catching equity is better with KQ/AQ, because you’ll improve to the effective nuts on cards that are more likely to be bluffed. These may be small differences, but it’s good to have reasons for your plays because it allows you to maintain good frequencies while finding edges where others don’t.
Kebab
Dearborn, Michigan is known for its fine selection of Middle Eastern food. You might sample the best kebab you’ve ever had in a hole-in-the-wall off of Michigan Avenue. Or, say, at a private poker game, whose participants generously refer to the venue as “the club,” when it is in fact a room with a poker table, a TV, and a few couches, adjoined by a kitchen. And if that kitchen is inhabited by a chef who turns out morsels of beef, chicken, or fish, delicately marinated and then broiled, served alongside rice pilaf and fresh tomatoes and green onions, and if you can pay with chips — well, you really can’t beat it.
Jim Cramer
In 2009 Jon Stewart brought CNBC financial analyst-cum-goofball Jim Cramer on The Daily Show for a now legendary takedown. I often think about this interview because I think it gets to the heart of what’s going on with people like Daniel Amen. At one point this exchange takes place:
STEWART: Isn’t there a problem with selling snake oil and labeling it as vitamin tonic and saying that it cures impetigo… Isn’t that the difficulty here?
CRAMER: I think that there are two kinds of people. People [who] come out and make good calls and bad calls that are financial professionals and there are people who say they only make good calls and they are liars. I try really hard to make as many good calls as I can.
Cramer seems hurt by the insinuation that he would deliberately peddle bad advice. The issue here is integrity of intention versus integrity of method. If you watch the full interview, you’ll question whether Cramer has any integrity whatsoever, but let’s give him the benefit of the doubt and assume the “calls” ventured on the show really represented his best guesses about what the market would do. By presenting those guesses as financial advice, he was still behaving unethically, because they weren’t very good guesses and there was no reason to believe they were. His method wasn’t adequate to ground his conclusions.
By all accounts Dr. Daniel Amen gives the impression in person of being a kind, caring individual. He claims 75% of his patients report improvement after receiving therapy, which may well be true. None of that changes the fact that the science doesn’t support his claims about brain scans.
I know people who believe that the Jim Cramers and Daniel Amens of the world are deceivers who know exactly what they’re doing. While that may be true of some of them, I don’t think it’s true of all of them, and I kind of doubt it’s true of most of them. It’s the rare person who’s cold-blooded enough to live with that image of himself. I suspect most of them believe good intentions are what counts. While this may be true in personal relationships, in some professional fields — including financial advice and psychiatry — intentions don’t count for much without methods.
Dusty “leatherass” Schmidt recently wrote a blog post by in which he announces his retirement from online poker on the advice of a psychiatrist named Dr. Daniel Amen.
After answering thousands of questions, having two SPECT scans of my brain (one resting and one concentrating) and undergoing numerous neurological exams, I received the news. Essentially what I learned was that I had developed a pattern in my brain (while concentrating) that facilitated a very high functioning processing loop. This “loop” was able to move so quickly and so efficiently that it was no longer a mystery why I was able to play poker at such a high level and across so many tables. I was told that as long as I was functioning “within the loop” I was going to excel at it. But if I had to do anything outside of the loop within a few hours of being in the loop for a long period of time, I would get “stuck.” When you’re “stuck” you can hear the other person speaking to you and you may even want to answer them, but the words or actions just never quite happen…
After hearing some exciting things about my brain, I then got the bad news. The bad news is that what poker puts the human brain through is unsustainable. I was told that if I choose to play online poker as a full time career, I will likely die of a stroke before the age of 50. They said that the human brain simply does not have the capacity to put itself through 8+ hour days of mass multi-table online poker. They said that playing online poker to that degree was quite literally one of the very worst things you can do to your brain and body. They said it wasn’t quite as bad as being an NFL player or someone who works full time around organic solvents, but it wasn’t far off. It was on the doctor’s urging that I retire from full time online poker.
I’m actually not that worried about this. Aside from the highly dubious nature of the diagnosis, my own laziness provides a rock solid safeguard against playing the kind of volume Schmidt is known for. While I do worry about how much time I spend thinking about a stupid game, I hardly need a psychiatrist to tell me why that might be a problem.
What I am interested in is the nature of evidence, belief, and fraud. When a friend told me about Schmidt’s blog post, I was immediately skeptical about the role of brain scanning in the diagnosis. My girlfriend, Ali, is a neuroscience grad student, and I know from talking with her that the use of brain images in research is still at best provisional. Our knowledge of the brain is not advanced enough to draw detailed conclusions from these types of images. In light of that it seemed extraordinary that a psychiatrist could make sweeping claims about what poker does to the brain, based on images obtained from one person. Furthermore, Schmidt admits in his blog post to problems with drugs and alcohol; wouldn’t that be a sensible first place to look when evaluating his issues with his family?
Schmidt’s paragraph on Dr. Amen reads curiously like an advertisement: “Dr. Amen is well known around the world as perhaps the most innovative and successful psychiatrist in the world…His Clinics are world class…His work is world renowned.” This isn’t the language of psychiatry or of a patient describing his doctor; it’s the language of self-promotion. It kind of sounds like Schmidt is repeating verbatim the pitch he got at the clinic.
A Google search on Dr. Amen turns up a slew of critical accounts. This Washington Post piece is perhaps the most thorough: the author was careful to source quotes from a variety of sources in the neuroscience community; I wasn’t counting, but I’d guess at least a dozen scientists are quoted in the article. None of them are remotely convinced of the validity of Dr. Amen’s methods and claims.
This is where, if I were writing for a magazine or newspaper, I’d say something hedging and politic like, “There are serious doubts about Dr. Amen’s work,” or, as in the Washington Post,
Amen’s career is very troubling, for one of two things must be true.
One, Daniel Gregory Amen, born in 1954 in Encino, Calif., son of Lebanese immigrants, is 20 years ahead of virtually the entire psychiatric field (he says about three dozen other clinics use SPECT scans, but few as profusely as he does), and the establishment has failed to recognize a historic breakthrough.
Or, two, the man has grown fabulously wealthy — he lives in a $4.8 million mansion overlooking the Pacific Ocean — by selling patients a high-priced service that has little scientific validity, yet no regulatory body has made a move to stop him.
But this is a blog, so let’s be real: Dr. Amen is full of shit. Clearly, this is also the point of view of the Washington Post article, but there seems to be a rule of journalism that the author must maintain an appearance of “balance” even if the evidence is predominantly on one side.
Something weird happens on the wikipedia pages of people like Dr. Amen. (By “people like Dr. Amen” I mean people who are almost certainly full of it, but who have not been unequivocally debunked, whether because their field is esoteric, they are not prominent enough to warrant the effort of discrediting them, or for other reasons.) The sites tend to be created and maintained by fans of the subject (or perhaps even the subject himself in some cases). As such, they give the impression that the subject is generally considered legitimate, with a whiff of controversy; when in fact he is generally considered a fraud, although there remains a shred of doubt. The talk side of the page often becomes an argument between passionate supporters and bemused wikipedia editors, who have no personal interest but stumbled on the page and noticed something fishy. I previously wrote about Andriy Slyusarchuk, who is a more straightforward — and hilarious: “The Head of Department for Public Affairs of MIA of Ukraine, Volodymyr Polishchuk said that ‘independent examination is conducted to check the mental health of Slyusarchuk, as well as his possible psychic and hypnotic abilities’” – example of this phenomenon than Dr. Amen.
I feel like there’s a lesson somewhere in here about our media and our understanding of truth. Self-promoters and bullshitters engage in their pursuit full-time, with all their energy. Meanwhile, it’s not in anyone’s interest to devote the necessary time or energy to straighten things out. Traditionally that may have been the purview of journalists, but these days few of them seem to honor that responsibility very scrupulously. That’s how someone like Daniel Amen becomes the most popular psychiatrist in America.
Poker and the Brain
New Article on Two Plus Two Magazine
New Article on Two Plus Two Magazine
I have an article in this month’s Two Plus Two Online Poker Magazine. And I’ll have some more posts up soon!
Guesstimating (2)
The two men in the pictures were craigslist CEO Jim Buckmaster and neuroscientist Paul Buckmaster. So far as I can tell, they are not related…although, if I had known their last name was Buckmaster, I would have thought it likely that they were. (How many Buckmasters can there be?!)
Two more along the same lines:
1. At a nine-handed poker table, what is the chance at least one player has committed a murder?
2. How many of these 196 “customer reviews” were written by Tyra Banks or representatives of Tyra Banks?
You’re Doing It Wrong: Limping
The online poker community is divided into “regs” and “fish.” Regs are professionals or semi-professionals and winning players. Fish are recreational and usually losing players. Clearly, it’s important to be able to distinguish between the two as quickly as possible. The best way to do this using stats is to look at the difference between VPIP and PFR.
VPIP means “voluntarily put money in the pot”; it’s basically the percentage of hands that you play. PFR means “pre-flop raise,” which is self-explanatory. Regs typically have their VPIP and PFR very close together, say 23/21. This means that if they’re playing the hand, they’re almost always raising. Fish often have numbers that are far apart, say 12/38. They tend to enter many pots by just calling.
While regs will sometimes call a raise, for the most part they never just call the blinds. This is known as limping in. If it’s folded to a reg and he’s going to play the hand, he’ll play it for a raise. This tendency is so strong that if I see a player limp in even a single time, I immediately mark him as a fish: a weak player and a target.
Yet when you start playing live poker, something strange happens: most players limp more often than they raise. This is true even of some players at higher stakes, and even some players who are generally regarded as skillful.
So what’s wrong with limping? The backbone of the theoretical case against limping has to do with fold equity. Whenever you bet there’s a chance that everyone folds and you take the pot. The value you get from potentially taking down the pot is called fold equity. If you raise, there’s a chance everyone will fold and you win the blinds. If you just limp in, there’s no way you can win the pot immediately. At the very least, the big blind has already paid enough to see a flop. The blinds may not seem like much, but in a tough game every little bit of value calls.
That’s a theoretical argument, but the reality of most live games, which tend to play loose-passive, is that a single raise rarely takes down the blinds. The more relevant arguments against limping have to do with deception and post-flop playability. In my experience a typical live poker player’s pre-flop strategy often looks something like this:
Premium hands (AA-KK, sometimes QQ and AK): sometimes raise, sometimes limp hoping to re-raise
Medium-strong hands (AQ, TT, etc.): raise
Speculative hands (76s, 22, etc.): limp and call a raise
This strategy seems to make sense: try to get as much money as possible in with the strongest hands, while trying to see a flop as cheaply as possible with speculative hands. Yet it turns out to be a very easy strategy to play against. When someone using this strategy limp-re-raises, we can be certain he has a premium hand and either fold with little invested, or call and try to “crack” his hand. For example, if we have KQo we can comfortably fold. If we have 44 and we’re deep enough we can call, knowing that we’re extremely likely to win his whole stack when we flop a set and we can fold otherwise.
When he raises, we can attack him somewhat liberally with re-raises, knowing that many of his strongest hands like AA and KK would have limped; or we can call and play well against him post-flop. For example, we know that very few hands in his raising range are going to be happy to see an all-low card flop like 754.
His raising and limp-re-raising ranges are too transparent, but it’s his limp-calling range that’s really going to cost him a lot of money. This range is made up entirely of weak/speculative hands and will most often end up playing out of position (any player not in the blinds who raises will have post-flop position), without initiative (the opponent has the betting lead). Any one of these factors in isolation — weak range, out of position, without initiative — make it very difficult to play profitably. Combined together they equal a conflagration. Against many players, in fact, it’s profitable to isolate (raise in position) and continuation bet every flop. In other words, someone who knows nothing about poker and hasn’t looked at his cards can automatically defeat this strategy just by raise-betting.
That seems like a bad spot to be in. If we want to incorporate limping into our game and avoid getting abused by good players, we’ll have to be more careful to balance our ranges. For example, we should sometimes limp-re-raise as a bluff, maybe with something like A5s, to balance the times we limp-re-raise with premium hands. And we should sometimes limp-call with very strong hands like KK to strengthen our post-flop range in those spots. In any case, it’s clear that to create a balanced and deceptive strategy involving limp-calling, limp-re-raising, and raising is going to be very tricky and require us to do careful “accounting” of our own ranges.
A much simpler solution — and in practice the solution adopted by most successful players — is to not limp at all. If we’re going to play the hand, we just raise. This gives us the initiative and keeps our range balanced and opaque. It’s far easier to create a single balanced range than to juggle many ranges at once.
There are some situations in which I like to limp. Basically, these boil down to extremely passive games or extremely aggressive games. If the game is very loose-passive (many players see a flop, but there isn’t much pre-flop raising) and many of the players are very weak post-flop, it can work out well to limp speculative hands like low pocket pairs, mostly hoping to flop a monster and win a huge pot. Conversely, if there is a player behind us who’s a maniac, raising 50% or more of hands, an argument can be made for limping any hand we’re going to play. Most of the time, he’ll raise and we get to see what the rest of the table does before reacting.
In any case, limping should be at most a “specialty” play for certain specific situations. Raising is the backbone of every tough player’s pre-flop strategy.
More Coup
A few more thoughts after playing some more…
I’ve introduced the game to several people and found that new players have very consistent tendencies. In general they’re too risk-averse, manifesting in
1. Being reluctant to call bluffs, and
2. Not bluffing enough. When they do bluff, they often exhibit extreme timing tells. That is, if they actually have it they will block instantly. If they don’t, they’ll think a long time before bluffing.
So if you’re playing a new player, you can get away with profligate bluffing. If you’re just starting — bluff more!
I still haven’t played with more than two players, looking forward to that.
