Thinking in Bets
“Thinking in Bets” by Annie Duke
Life is more like poker than chess. In chess, the pieces are known, piece positions are known, movement patterns are known, and you can see them at all times. At its extreme, chess is like a large mathematical exercise where there is always an optimal move. However, in poker, much like life, information is hidden, opponents can purposely deceive you, and a lucky hand can still beat a strategically played hand. At anytime, a new poker player still has a chance at winning the pot. But in chess, a new player will almost always be destroyed by a more seasoned player.
In light of this, we should “think in bets”. That is, we should weigh outcome and decision quality separately, find truth seeking groups that purposefully engage and express opposing viewpoints to find the truth, and practice sound, scenario planning.
- Outcomes can be bad despite sound, favorable decision making
- Humans are naturally biased towards believing that bad outcomes are the results of bad luck or an unlucky situation; similarly, humans are biased towards believing that good outcomes are only the result of superior intelligence, planning, etc.; TL;DR - we’d like to think we the most control over our outcomes
- The same bias is true, in reverse, when we analyze others choices; good outcomes are the result of favorable position and luck; bad outcomes must be because bad decision making, incompetence, poor planning, etc
- Life is more like poker than chess; in chess, at the highest level, the results are largely based on superior probabilistic reasoning, devoid of luck — the pieces are always the same, they move the same, and there is little randomization; however, in poker, like life, elements of luck, deception, and unknown information are part of every decision that is made
- If we want to learn from our mistakes we need to try to consider the factors which lead to an outcome — how much of it was luck? How much of it was not knowing or considering all the facts? Did we make a good bet and get bad luck? Did we make a bad bet and have good luck?
- As a poker player increases their skill, they become adept at knowing all the odds (as best they can) before they make a decision, and they consider the fact that luck may always trump what was sound decision making
- It is a fallacy, and a unwise choice, to curb our behavior simply because of a bad outcome; if our decision making was sound, like Pete Carrol choosing to throw a pass at the end of Super Bowl XLIX, then we’re okay; we can keep making that decision with confidence that more often than not, the odds will be favorable
- Our outcome analysis, when we’ll considered, should most likely fall somewhere between how we’d analyze the outcome for ourself (and its biases) and how we’d analyze it for someone else (and its biases)
- To eliminate bias in truth seeking groups, discuss your decision making process while leaving out the results and your beliefs
- Try to take emotions and bias out of truth seeking
- For example, in the past, US Supreme Court justices used to appoint several underlings on the other side of the political spectrum to hear their opposing view points so they could make a more sound ruling
- Use the 10 minutes/10 days/10 years system for analyzing decision and outcome quality
- Consider your past and future self when making decision (“night Jerry Sienfield making decisions to stay up, hurts morning Jerry productivity”)
- Ulyssess contract: To prevent yourself from doing something you don’t want, you agree to a contract that will prevent you from doing that thing (ex: tie my arm behind my back so that I cannot try to scratch my face wound even when it itches)
- Scenario planning: Think about all possible futures, the probability that they will happen, and then take the best option
- What is better than scenario planning (future forecasting) is working backwards: think of the best possible future and walk it backwards to determine what would have to happen for that future to occur (ex: I know when our app is launched, we’ll need to be able to quickly adjust user flows, and swap-in new screens, so if that app was already here right now, then how is it built such that it enables that level of flexibility
- Combine back-casting (positive futures) with pre-mortems (negative futures; pre-mortems task us with looking at a bad outcome and determining “how we get here”
- When you do both, you more accurately represent all the possible futures and consider factors often overlooked when you only consider the “happy path”
- The future is always uncertain. Accept it. Calculate the odds. Consider the unbiased facts. Make a bet. Then rinse and repeat without seeing your past (the wins and losses) through a biased lens. As long as you keep making the best bets, that’s all you could hope for in a world of uncertainty.