do things with asymmetrical rewards

the single best heuristic i have for deciding whether to do something is: what’s the worst case, what’s the best case, and how far apart are they?

most of the things i’m glad i did are the same shape. i emailed a researcher at md anderson as a high-school sophomore asking if i could help on a cancer-classification project. worst case: i never heard back and lost the twenty minutes it took to write a polite email. best case: i did hear back, and i spent the next eight months learning how to use pytorch on real medical data, which is how i ended up with a line on a resume that opened most of the doors i’ve walked through since. one of those outcomes is bounded by the time to write an email. the other isn’t bounded by anything.

cold-applying to a quant firm with no track record was the same shape. so was applying to turing scholars. so was sitting down for my first usaco contest, which was three hours of a weekend against a medal that keeps re-opening doors years later. once you start looking, the pattern is everywhere: capped downside, long right tail.

the reason people pass on these isn’t that the math doesn’t work. the math is obvious the moment you draw it. people pass because the downside, while small in dollars or hours, is socially embarrassing. sending an email no one answers feels like a tiny rejection. asking a question in lecture and getting it wrong feels like a tiny rejection. applying to a thing and not getting it feels like a tiny rejection. these costs are real, but they’re bounded by something like ten seconds of awkwardness. the right tail isn’t bounded by anything.

a loose rule i try to follow: when i notice i’m talking myself out of something because i’d look silly if it didn’t work, i make myself do it anyway. it’s almost always a sign i’m mispricing the downside.

there’s a version of this that gets too aggressive, where you take every long-shot bet you see and your life turns into spray-and-pray. what i actually care about is narrower. when a specific opportunity has the asymmetric shape and i notice i’m talking myself out of it on social grounds, i take it. the cost of saying yes too often to those is small. the cost of saying no too often is, eventually, your life.

you don’t have to be lucky very often. you have to be positionedfor luck when it happens. the way you position yourself is by taking enough of these bets that something eventually hits, and then making sure you’re the one holding the rope when it does.