Thinking in Bets: a poker-master's Jedi mind-trick for being less wrong

Originally published at:

Annie Duke dropped out of a PhD in cognitive psychology to become a professional poker player; now she runs a nonprofit devoted to improving decision quality by merging the practical cognitive tools of the world’s greatest poker players with the leading edge of cognitive psychology, a method she describes in an excellent and charming new book called Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts.


Great review and nice caveats.

I wish books like this were like some Netflix series, sometimes it only needs to be 6 episodes. Don’t stretch it out to 12.


It’s always worth taking lessons from the losers, if you have the wit to see what the lessons really are. Sepp Herberger, prominent supporter of the Nazi party and manager of the German national football team 1921-1925 summed the whole thing up quite well:

“Nach dem Spiel ist vor dem Spiel”
(After the game is before the game)


If you’re able to just choose what your self-worth depends on, it doesn’t seem like you’d need self-help books…


Thanks to this revolutionary technique, I can stop feeling like a loser just because none of my plans worked out. Now, I can feel like a loser because I had shitty plans and executed them poorly!

I’m just kidding. Welllllll… I’m sort of kidding.


Thinking fast and slow is still my bible for cognitive biases. The real trick is not finding one bias to beat like this woman but rather keeping them all rationally weighted in our attention. Bayesian thinking is what ties this all together: weighting new info like this in proportion to how much it really affects outcomes.


This reminds me of the annual predictions of the author of the Star Slate Codex blog.

At the beginning of the year, he makes a lot of predictions, and gives what he considers the chance of each one coming true. And the end of the year, he goes through each one, ranks whether he was accurate, and then compares the totals with predictions of the same probability.

He then checks whether his accuracy was below or above what he predicted, and the important part, being above is just as “bad” as being below.

If 9 out of 10 of you “50% likely” events come true, that’s less accurate than if only two of them of them were accurate.

That’s the only place I’ve ever seen do that, and one that I’d love to see professional prognosticators do.

The whole “I think X, Y and Z will happen” type prediction is utterly useless except from an entertainment perspective and it often worries me that the people who are making actual decisions are so rarely able to comprehend the basics of probability and statistics.


Where this idea breaks down for me, is how formal gambling takes place within a tightly constrained set of circumstances. Where in the real world…

I mean, even among my friends who accept human caused global warming as real, I can’t seem to find anyone willing to change their stance on driving because of that.

If everyone is wrong about the state of the game, then it doesn’t matter what we think the odds are, history is going g to surprise us.


Annie makes a huge presumption about poker. A lot of poker is making other people think you have a good hand when you don’t, or hiding a good hand when you do. She makes it all about your hand and deciding what to do. I never play my hand. I always play the opponent’s hand.


“[…] mediocre poker players, and other normies […]”

Seriously ?


Finally, there’s an unsupported implication of universality in the method proposed. The tactics Duke catalogs have worked for many people, but it’s not clear whether they’d work for everyone.

This x a million. Any self-help book that doesn’t start with the words “this worked for me, but that doesn’t mean it will work for you” should be thrown aside with great force, even if the contents are generally worthwhile.
(Having said that, I may still give this one a look as it sounds as though it’s more coherent than most.)

1 Like

Players like Annie Duke make a lot of money playing against players like you.


You would think, but she doesn’t.

1 Like

Isn’t that a dissonance in averred versus actual values, rather than a misunderstanding of the outcomes of their actions?

1 Like

As far as betting on a future goes, that’s a distinction without a difference. We aren’t playing the same game, we’re not even in the same casino.

She describes how good poker players separate “good decisions” from “good outcomes.” A decision is good if it gave you the highest chance of winning based on the information you had, even if you ended up losing. Mediocre poker players and other normies suffer from “resulting,” in which results, not information and reasoning, are used to evaluate decisions

To point out the obvious, this is only interesting for making relatively simple decisions in which you know the odds of each outcome. Like poker. If high level human decisions were like that, they’d be easy to make.
If you don’t know the odds of something analytically, what she dismisses as “resulting” is the way that you figure out how to estimate those odds from the results.


I don’t have any experience playing professional poker, but I have some experience with friendly games. As you point out, there is the poker process where you calculate the odds of your own hand winning against other potential hands by using the visible cards. And then there is the psychological and observational process where you try to understand HOW other players think, and what motivates them in each situation. It is entirely possible to be good at one of these processes or the other. But to be really good at poker, you have to be good at both, and you also have to understand yourself in the way that others view you.

1 Like

I think you might be confusing acceptance of the facts with having the same utility function you do.

It is very possible (and mostly likely) to accept the same set of facts that you do, clearly understand the trade-offs, and and make an entirely different decision.

And as soon as you turn such decisions into “well, if you understand the facts and choose differently, then you are a bad person”, then you’re incentivizing creating facts in order to justify one’s decision, and we all do enough of that, thank you very much.


“Goes to show, you don’t ever know,
Watch each card you play and play it slow”

I know nothing about poker, and even I can tell this is a flawed comparison, but just for fun I tried out my google-foo. I discovered, where there are lots of charts that look like this:
I’m guessing they look this way because you have better things to do and don’t play as much as she used to?

1 Like