Thinking in Bets

“Wanna bet?” redux

Annie Duke calls the process of assigning causes to results outcome fielding. Treating that explanation as a bet can help us examine a habitual story about luck and skill. If someone challenged us to a meaningful bet on our explanation, we might scrutinize it more closely. If we wanted to win that bet, we wouldn’t reflexively field bad outcomes as all luck or good ones as all skill. (If you walked into a poker room and threw around words like “always” and “never,” you’d soon find yourself challenged to a bunch of bets. It’s easy to win a bet against someone who takes extreme positions.)

Imagine getting into an accident at an intersection after losing control of your car on a patch of ice you couldn’t see. Your first thought would likely be that you got unlucky. But what if you had to bet on that? Depending on the details, there would be a lot of alternatives to just plain bad luck you would now consider. Based on the weather, maybe you could have anticipated some ice on the road. Maybe you were driving too fast for the weather conditions. Once the car started sliding, maybe you could have steered differently or maybe you pumped the brakes when you shouldn’t have. Maybe you could have taken a safer route, choosing a main road that would have been salted. Maybe you should have kept the Mustang in the garage and taken the Suburban.

Some of the reasons we come up with may be easy to discount. And some may not. The key is that in explicitly recognizing that the way we field an outcome is a bet, we consider a greater number of alternative causes more seriously than we otherwise would have. That is truthseeking. Duke cites poker player Phil Ivey as an example of this habit.

The prospect of a bet makes us examine and refine our beliefs, in this case the belief about whether luck or skill was the main influence in the way things turned out. Betting on what we believe makes us take a closer look by making explicit what is already implicit: we have a great deal at risk in assessing why anything turned out the way it did. That sure sounds like a bet worth taking seriously.

Treating an explanation as a bet gives us a reason to test it: being accurate matters more than telling a flattering story. Duke proposes using the satisfaction of learning as a reward for this practice. Repetition can help make it a habit, although wanting to win can also tempt us to defend a favored explanation.

Overconfidence has been called the "mother of all biases." A contrast between 100% confidence and 80% accuracy illustrates the problem, but these figures are not a universal rate. A "betting" mindset asks us to specify confidence and compare it with outcomes over time. A confident declaration might be "I'm sure I'm right." Asking what odds we would accept can expose how much uncertainty remains. Calibration means matching stated probabilities to observed frequencies across comparable predictions; merely hedging does not establish it.

Thinking in bets can prompt exploration of alternative hypotheses, including reasons supporting conclusions opposite to our habitual self-serving account. We are more likely to explore the opposite side of an argument more often and more seriously—and that will move us closer to the truth of the matter. Duke’s approach encourages the kind of review useful in forecasting. They don't just say "I was unlucky" when a prediction fails or "I'm brilliant" when it succeeds. They ask, "Was I right for the right reasons? Or did I get lucky when a low-probability event happened to work in my favor?" They treat every outcome as an opportunity to examine their reasoning, identifying both what they got right and what they got wrong. Reviewing successes as well as failures can expose mistakes that a favorable result concealed.

Thinking in bets also triggers perspective taking, leveraging the difference between how we field our own outcomes versus others’ outcomes to get closer to the objective truth. We know we tend to discount the success of our peers and place responsibility firmly on their shoulders for their failures. A good strategy for figuring out which way to bet would be to imagine if that outcome had happened to us. If a competitor closes a big sale, we know about our tendency to discount their skill. But if we imagine that we had been the one who closed the sale, we are more likely to find the things to give them credit for, that they did well and that we can learn from. Likewise, when we close the big sale, let’s spare a little of the self-congratulations and, instead, examine that great result the way we’d examine it if it happened to someone else. We’ll be more likely to find the things we could have done even better and identify those factors that we had no control over. Perspective taking can reveal an inconsistency between the standards we apply to ourselves and to others. The better explanation is the one supported by the evidence, wherever it falls; it need not lie halfway between the two judgments.

Survivorship bias offers another reason to examine the comparison. Successful people are visible "survivors"; similarly promising failures may be absent from view. Trying to "visit the graves" means looking for those missing cases before explaining success solely through talent. We cannot assume that we "beat" everyone through skill, nor that every unsuccessful person was equally capable. Comparing the cases helps test both claims.

Once we start actively training ourselves in testing alternative hypotheses and perspective taking, it becomes clear that outcomes are rarely 100% luck or 100% skill. Luck and skill can both contribute, and their roles vary between cases. Considering both avoids forcing every result into a single category. This means that when new information comes in, we have options beyond unquestioned confirmation or reversal. We can modify our beliefs along a spectrum because we know it is a spectrum, not a choice between opposites without middle ground.

This makes us more compassionate, both toward ourselves and others. The psychologist Paul Rozin observed that negativity trumps positivity with force—a single cockroach ruins a bowl of cherries, but a cherry does nothing for a bowl of cockroaches. The image conveys how a negative experience can outweigh a positive one. It does not establish a fixed ratio between the feelings of losing and winning or of being wrong and right. We are in a better place when we don’t have to live at the edges. Euphoria or misery, with no choices in between, is not a very self-compassionate way to live.

Treating outcome fielding as bets constantly reminds us outcomes are rarely attributable to a single cause and there is almost always uncertainty in figuring out the various causes. Identifying a negative outcome doesn’t have the same personal sting if you turn it into a positive by finding things to learn from it. You don’t have to be on the defensive side of every negative outcome because you can recognize, in addition to things you can improve, things you did well and things outside your control. You realize that not knowing is okay.

Certainly, in exchange for losing the fear of taking blame for bad outcomes, you also lose the unadulterated high of claiming good outcomes were 100% skill. That’s a trade you should take.

You also become more compassionate toward other people when you treat fielding outcomes as bets. When you look at the outcomes of others from their perspective, you have to ask yourself, “What if that had happened to me?” You come up with more compassionate assessments of other people where bad things aren’t always their fault and good things aren’t always luck. You are more likely to walk in their shoes. Imagine how Bartman’s life would have changed if more people worked to think this way.

These judgments also shape the stories we tell about ourselves. Recognizing that an outcome may combine choices and circumstances can make that story less dependent on blame or self-congratulation. Saying that an outcome was rarely 100% one or the other leaves room for uncertainty; it does not establish a precise division between luck and skill.

The hard way

Thinking in bets is hard, especially initially. It has to start as a deliberative, System 2 process, and will feel clunky, weird, and slow. Our default, automatic System 1 response is to be overconfident in our self-serving narratives. The deliberate act of imagining a "bet" on that narrative is a System 2 intervention, forcing calibration and reducing overconfidence. Certainly, there will be times it doesn’t make sense. Like if you don’t get a promotion at work, you’re probably going to wonder how you’re supposed to feel better acknowledging that so-and-so was more deserving and that you could learn a lot from them. It takes work to avoid the temptation to blame it on the boss being a jerk who doesn’t know how to evaluate talent.

That feeling is natural. I built my poker career out of these principles of learning and truthseeking, yet I still catch myself falling into the self-serving bias and motivated reasoning traps. Reshaping a habit requires time, preparation, practice, and repetition.

Look at other kinds of habit changes. If I had a habit of getting out of bed at midnight to eat a cookie, it takes work as well as will to change that habit. I have to identify the habit I want to change, figure out the routine to substitute, and practice that routine in deliberative mind until the habit is reshaped. I would need to stock apples in the house, and keep them more readily available than cookies. Then I need to actually eat the apple at midnight instead of reaching for the cookie, repeating this routine until it becomes a new habit. That takes work and willpower and time.

Striving for accuracy through probabilistic thinking takes practice. It may become easier with repetition. We have to start doing this with deliberation and effort, but it eventually becomes a habit of mind. Just like declaring uncertainty in your beliefs, it eventually goes from a somewhat goofy and awkward extra step to a habit integral to how you view the world around you. The goal is a better-calibrated account, with confidence that reflects the evidence. Sometimes that requires greater confidence; sometimes less.

To be sure, thinking in bets is not a miracle cure. Thinking in bets won’t make self-serving bias disappear or motivated reasoning vanish into thin air. It can still help us notice learning opportunities we would otherwise miss. The value depends on what we learn and how we use it.

Poker’s compressed version of real-world decision-making showed me how being a little better at decision-making could make a big difference. A poker game can consist of a few hundred hands. Every hand can require up to twenty decisions. If, over the course of a game, there were a hundred outcomes that provided learning opportunities and we caught ten of them, we would still be missing 90% of our chances to learn. We wouldn’t have transcended the way our brains function and built ourselves a different brain, but we don’t need to. If our opponents are people like Nick the Greek, they are missing almost every learning opportunity. We’re obviously going to do better than they do just catching 10%. If another of our opponents is someone just like us, but who isn’t working to transform their outcome-processing routines, maybe they (this prior version of ourselves) will pick up five opportunities. Again, we are missing 90%, and we are still going to clean up on an opponent who is trying to learn but doesn’t know how.

The benefits of recognizing just a few extra learning opportunities compound over time. The cumulative effect of being a little better at decision-making, like compounding interest, can have huge effects in the long run on everything that we do. When we catch that extra occasional learning opportunity, it puts us in a better position for future opportunities of the same type. A useful lesson can improve later decisions of the same kind. Think of it like a ship sailing from New York to London. If the ship’s navigator introduces a one-degree navigation error, it would start off as barely noticeable. Unchecked, however, the ship would veer farther and farther off course and would miss London by miles, as the distance traveled increases. This is an analogy for noticing and correcting an error, rather than a mathematical example of compound growth.

The first step is identifying the habit of mind that we want to reshape and how to reshape it. That first step is hard and takes time and effort and a lot of missteps along the way. So the second step is recognizing that it is easier to make these changes if we aren’t alone in the process. A partner can help us notice explanations we would otherwise accept without question.