You gotta know when to hold 'em, know when to fold 'em

May 3, 2018 9:10 AM

A little more than a year ago, New Yorker writer Maria Konnikova announced that she was diving into the world of professional poker as a new player, all for the purpose of writing a book about her experiences. Yesterday, PokerNews reported that the actual writing of the book is on hold because Konnikova, under the guidance of pro Erik Seidel, got too good at poker.
posted by Chrysostom (32 comments total)

On the plus side, I guarantee you someone's already half through writing the script for this movie.
posted by selfnoise at 9:12 AM on May 3, 2018 [10 favorites]

Konnikova previously wrote a book on the psychology of confidence games, so I wonder how applicable/helpful that might be to swimming with the sharks in pro poker.
posted by Strange Interlude at 9:23 AM on May 3, 2018 [9 favorites]

...Konnikova decided she had to push the book schedule back and go all in on poker for the time being.

I spotted your tell, there.
posted by Greg_Ace at 10:57 AM on May 3, 2018 [6 favorites]

Doesn't this happen fairly often, getting profitably good at poker? Like it can be learned? I wonder if there's a success bias toward newbies, since nobody knows their tells yet.
posted by rhizome at 11:13 AM on May 3, 2018 [2 favorites]

Competitive poker has got to be flush with Dunning–Kruger. I'm not very smart and it seems utterly unpleasant and dull, really smart folks I'd expect find a way to make a good living doing something they enjoy.
posted by sammyo at 11:16 AM on May 3, 2018 [1 favorite]

Doesn't this happen fairly often, getting profitably good at poker? Like it can be learned? I wonder if there's a success bias toward newbies, since nobody knows their tells yet.

I think it's more likely to be survivor bias. Most new players crash and burn and lose money and then quit. The lucky ones think they're poker gods.

For a while.
posted by RustyBrooks at 11:18 AM on May 3, 2018 [2 favorites]

Win or lose the resulting book/movie should be pretty interesting. You can bet on it.
posted by tommasz at 11:23 AM on May 3, 2018 [2 favorites]

A few years ago, I saw an article about a group of guys who were pro Magic: The Gathering players. They made some very nice cash doing that, but it's time intensive, requires some investment of cash, and the pay isn't THAT good. At least not compared to poker.

So they retired from Magic and used the card theories from that game to learn poker and ended up like many many times millionaires.

I'm certainly not a professional magic player, but I am at least better than average, and I do have a very deep understanding of the theories behind being competitive.

I've often fantasized about trying what those guys did.
posted by Twain Device at 11:29 AM on May 3, 2018 [1 favorite]

Competitive poker has got to be flush with Dunning–Kruger.

All gambling lives and dies by Dunning-Kruger, but you can indeed learn to maximize your statistical chances of winning in poker at any given moment, and to read some or all of your opponents to some degree.

The people who win the big tournaments these days often are much less good than they think they are. Basically any one person playing an ultra-aggressive style will be wiped out early, but, when you put 10,000 people playing an ultra-aggressive style in a tournament, one of them is bound to ride the wave of stupid luck required to a victory. But consistent success as a professional playing against other professionals requires quite a bit of skill.
posted by praemunire at 11:31 AM on May 3, 2018 [6 favorites]

I just love this. I'm sure there are many reasons why this could be a fluke, or that poker is something people can do well for a while and then sputter out at, but I love that book research could turn into high-stakes big money winning for someone.
posted by xingcat at 11:52 AM on May 3, 2018 [5 favorites]

So they retired from Magic and used the card theories from that game to learn poker and ended up like many many times millionaires.

M:TG is both so broad and so deep that I think for a player to be "very skilled" at it implies that they have acquired transferrable skills. It's got to be brain training of a kind.

I'm a hobbyist board game designer and kind of a crap M:TG player. But every time an actual M:TG player playtests a game of mine, they provide the best feedback. Sometimes just after the rules are explained, before the first turn has started.
posted by a snickering nuthatch at 11:54 AM on May 3, 2018 [12 favorites]

I hadn't heard that about M:TG players, but in the dawn of online poker, there were a lot of former competitive Starcraft players, and those guys pioneered the practice of running many many poker tables at once, because they were people with a proven track record of making many decision per minute following multiple parallel strategies. Some of these guys would play 20 tables at a time. With a lot of effort I could play 8 and without too much effort I could do 4.
posted by RustyBrooks at 12:06 PM on May 3, 2018 [2 favorites]

xingcat: but I love that book research could turn into high-stakes big money winning for someone.

If only this were transferable to other subjects.
posted by clawsoon at 12:41 PM on May 3, 2018 [1 favorite]

A few years ago, I saw an article about a group of guys who were pro Magic: The Gathering players. They made some very nice cash doing that, but it's time intensive, requires some investment of cash, and the pay isn't THAT good. At least not compared to poker.

So they retired from Magic and used the card theories from that game to learn poker and ended up like many many times millionaires.

Yeah in my teens I was into MTG enough to be able to name pro players (and in fact some local people I was acquainted with ended up becoming pro players a little later) and I think when poker blew up a significant chunk of that community migrated to it, at least for a little while.

(More than a couple of them also ended up in fintech or assorted startup shit.)
posted by atoxyl at 12:43 PM on May 3, 2018

Magic is a way more interesting game to me than poker, but as far as I know most of the prizes for top players are provided by WotC/Hasbro itself, presumably as a means to encourage lesser players to spend money on the game. So it's really a very small number of people who can make any sort of living on it at a given time.

edit: top lifetime winnings
posted by atoxyl at 12:48 PM on May 3, 2018 [1 favorite]

I used to play poker. A lot. I was not in a relationship and I had nothing but free time. I would keep spare clothes at work in case I didn't have time to go home. I would go to the casino and play as soon as I got off work, and sometimes play until the next morning. I generally played a 2-4 or 4-8 tables (the blinds). After doing this for about a year I decided to actually track what I was making, and I was making money, because I only played out of my roll, and it was getting bigger. I'm a fairly analytical guy, but it took some time before I even thought to apply metrics to my game.

I loved the table banter. College kid: I'm here to earn enough for tuition. Me: I'm just here to take that stack of chips in front of you.

I had several times were someone at the table nearly got violent, and at least once where I asked for a security escort to the car.

I loved playing. I paid for two transmissions for my car out of winnings, and still had money to play with. It was great! Free money.

But then I totaled up my winnings and divided this by the number of hours in the chair, and I figured out I was making $6 an hour on average. My friend told me, "I'd love to make $6 to play a game I enjoyed," and for a while this kept me going, but once I realize most retail jobs in my area were paying twice that, and my current job was paying five times as much, well, the fun went out.

I tried stepping up my game, sitting at the tables with the bigger blinds, but I tighten up so much when I am sitting a 16-32 table that I am way too easy to read.

When you sit a table with ten strangers who are all trying to take your money, it's a weird experience. You quickly identify the players who are better than you. You also know fairly fast those who are worse, and those who shouldn't be sitting there at all. If you stay out of the better players' pots, and only push them when you've got the nuts, you can generally all agree to fleece the others at the table. You bluff the players who are as good as you, crush the bad players, and capitulate to those who are better. Generally anyone who wants to can make money playing poker (assuming you have some semblance of a mind for statistics and probability mixed with risk-taking). You just have to be willing to put in the time and have to have a weird level of patience. The real challenge is figuring out exactly how good you actually are, as opposed to how good you think you are.

I miss poker, but I like having the time back.
posted by cjorgensen at 1:05 PM on May 3, 2018 [47 favorites]

If you’re interested in this kind of thing, check out Victoria Coren Mitchell’s excellent poker-themed memoir, For Richer, For Poorer which goes all the way from playing with her older brother’s friends as a teenager to her first European Poker Tour win.
posted by tomcooke at 1:22 PM on May 3, 2018 [4 favorites]

I'm a hobbyist board game designer and kind of a crap M:TG player. But every time an actual M:TG player playtests a game of mine, they provide the best feedback. Sometimes just after the rules are explained, before the first turn has started.

They understand how to break games inherently because MtG is a game that people often break. I dabbled in the Pro Tour before life took me in other directions. The best-of-the-best at it are generally extremely intelligent emotionally and intellectually but just kind of understand how any system flows. Magic is an extremely complex system that's always changing (new sets of cards, new "meta" where popularity of decks and cards shift, etc.) so the best at it are just... the best at mental games. (Mental as in there is no physical component, not as in hokey "mind games" though those are often important.)

Now the people right below the best players tend to be just filled with bile and rage so... yeah.

After much thought I can distill the essence of what makes someone good at a game to one idea: they win the games they should win, they win a decent amount of the games that are a tossup and they win way more than they should the games that any other player would lose.
posted by OnTheLastCastle at 1:23 PM on May 3, 2018 [4 favorites]

I am a former semi-professional poker player who did a couple years on the WPT circuit and I've played in every WSOP for the past 10 years although my bread and butter has always been cash games. I play very little these days, but I do still invest in backing some tourney pros (paying their fees in exchange for a percentage of their winnings). I do think that most intelligent people could learn to be profitable at poker if sufficiently motivated and exposed to the right education. People from gaming backgrounds often learn very quickly -- many top professionals have come from MtG and some from backgammon and chess and even strategy video games like Starcraft.

Eric Seidel is as brilliant a thinker about poker as anyone I've ever met and I have no doubt that being coached by him is worth a lot in terms of improving your game. But it is very difficult to judge the skill of tournament players by a relatively few results like Maria has so far, because luck plays a major part in your results in the short term. For instance, in her big win she had all her chips in the middle with 77 against another player's AA pre-flop and she caught a straight to double up. 80% of the time, she is out of the event and doesn't have the same fantastic story. She played the hand perfectly and it was just one of those things, but it illustrates a bit of how luck in the short term changes everything. I believe she fired two bullets in the Monte Carlo main event a few days and also got bounced from the high roller, so that's a loss of a bit over $40,000 in the past week. Outsiders rarely have the whole story.

I lack the temperament to handle the swings of tourney life and found I could make a lot more money on a consistent basis by playing cash games. I'm certainly biased, but I think successful cash game players usually have a more stable income. I invest in some extraordinarily talented tourney pros with lifetime winnings into the seven figures, who are essentially broke. Its a very common story.
posted by Lame_username at 1:53 PM on May 3, 2018 [20 favorites]

See also Victoria Coren (Mitchell) . I read her poker book last month, and it's terrific. Here she is writing on Molly's Game.
posted by Paul Slade at 3:10 PM on May 3, 2018 [1 favorite]

Whoops! Sorry, tomcooke - just realised you beat me to it. Good book, though, isn't it?
posted by Paul Slade at 3:11 PM on May 3, 2018

WHY DO I CHOOSE THIS FOR A LIVING?
posted by Mr.Encyclopedia at 3:21 PM on May 3, 2018

Unfortunately for Konnikova, the best thing she can hope for now is to die in her sleep
posted by rollick at 4:24 AM on May 4, 2018 [1 favorite]

Her appearances on The Gist for the segment, “Is this bullshit?” are so great!
posted by amanda at 6:19 AM on May 4, 2018 [1 favorite]

“6 Reasons Professional Poker Is Way Harder Than It Looks.” [article @Cracked.com/]
posted by LeLiLo at 12:32 AM on May 5, 2018

at times i've been in a weird social milieu where not gambling is seen as shameful. some people see poker as a way to practice achieving a good outcome by reasoning under uncertainty.

which of course it is, sort of. i guess like sports are a way to hone certain skills through play. but being bad at poker isn't the same as being physically out of shape or unskilled. i resent the idea that it is morally better to get good at gambling games rather than just refusing to sit down at the table.

it makes me wish i could adopt some kind of religion or code of conduct that just forbids gambling altogether.
posted by vogon_poet at 4:18 PM on May 5, 2018

In a different area of professional gambling, this article by Kit Chellel for Bloomberg about a pro gambler who built a statistical model to bet on horse races was well-written and fascinating.
posted by mbrubeck at 3:48 PM on May 8, 2018 [2 favorites]

> mbrubeck:
"In a different area of professional gambling, this article by Kit Chellel for Bloomberg about a pro gambler who built a statistical model to bet on horse races was well-written and fascinating."

Wild, somewhere in my archives is an OMNI magazine with a story about a guy who was applying in-your-head math to racehorse betting. This would have been around 1987, and since my hometown in the site of the county fair, and our fairgrounds has a race track, I tried to try it that summer! I am not good at math.

I wonder if it's the same guy. OMNI archives on the web don't seem to exist anymore, and it's not in a normal web search either.
posted by rhizome at 5:32 PM on May 8, 2018

If you're willing to put some work into it, there's a site with the tables of contents for OMNI.
posted by Chrysostom at 6:34 PM on May 8, 2018

wow, that is a weird article!

apparently the algorithm that got him started was literally just logistic regression (?!?) which is only a tiny bit more complicated than stats-101 linear regression. this is so basic i have to think he may be lying, but then again maybe not. i guess computers weren't widespread in the 70s to allow solving things numerically, but it's still weird to think about.

presumably the bookies have gotten wise to that now, and have their own more sophisticated machine learning models to set odds.
posted by vogon_poet at 9:50 AM on May 11, 2018

Bookies don't set the odds for horse racing. That's part of how this scheme works. It's parimutuel betting. You're not competing against the house, or the track. You're competing against other bettors. Note the bit in the article where the track calls them up and asks how they can help.

Machine learning models are great when you've got tonnes and tonnes of data. You can just pour it into an ML system and it will produce a model that provides some predictiveness, if there is predictiveness to be had. For smaller datasets, a trained statistician can produce better results using more traditional techniques, such as logistic regression. Traditional techniques can be more interpretable too. E.g., they can tell you that -- to contrive an example -- an extra day of rest for a horse makes them 0.8 seconds faster over a mile. A hammer is a lot less sophisticated than a 3D printer, but a hammer is still better at driving nails.

Also, they're making money at this not just because of statistical smarts. They had the inclination and the resources to gather data. They have full time employees doing data entry on the historical performance of thousands of horses. Not everyone can do that.

You also need a lot of capital to do this. It doesn't say this in the article, but it is certain that their edge is small. It takes a lot of money to exploit a small advantage in gambling.
posted by chrchr at 2:56 PM on May 11, 2018 [1 favorite]

i mean i'd count logistic regression as machine learning. and i guess as you say, part of the advantage is the interpretability, because he had to come up with a simplified approximation that he could do in his head.

it makes intuitive sense to me that whatever crazy operation they've got going in the new millennium would make money.

but the idea that one guy could get some data, and do essentially the bare minimum of statistics, and somehow still get an edge, is very surprising even for the 70s. it makes a little more sense though now that you point out that it's parimutuel which I hadn't thought about.
posted by vogon_poet at 3:04 PM on May 11, 2018