The faulty math of Sam Bankman-Fried's Earth destroying coin flip

SBF's accidental destruction of the planet Caroline Ellison testified Sam Bankman-Fried "would be happy to flip a coin if it came up tails and the world was destroyed, as long as it came up heads the world would be like more than twice as good."

This has been used to illustrate his high risk appetite as well as his willingness to bend and break rules in the name of the (probabilistic) greater good.

Most people would have issues with taking a bet where one of the possible outcomes is the destruction of the planet, but let's put that aside and analyze it purely on its mathematical merits.

As it turns out, math also says it's the wrong thing to do.

Imagine you take the bet and you win. The world wasn't destroyed and prosperity is now more than twice what it was. Great. But here's the question: would you be willing to take the bet again?

Nothing has really changed, other than the baseline being more than twice what it was. So if taking the bet the first time was the right decision, you should be willing to take it again.

But what happens if you keep winning and keep taking the bet again? Well, you're guaranteed you'll lose the bet at some point. So SBF thought a bet that has a 100% chance of destroying the Earth was worth taking. Maybe it shouldn't come as a surprise that his empire came crashing down.

What would it look like to apply the Kelly criterion to this bet? It might look like betting on a fraction of the planet to either get destroyed or more than double its prosperity, and picking the fraction so that the overall prosperity growth rate is maximized.

Most people would still have issues with risking a fraction of the world being destroyed, but the point is if you're going to justify your actions purely on the math, you should at least get your math right.