3 ms·

How do you reconcile this > hope that they could Martingale their way out of it with this > their JS background giving them bad instincts Surely, the instin

How do you reconcile this > hope that they could Martingale their way out of it with this > their JS background giving them bad instincts Surely, the instincts they got from JS is that you almost surely can't Martingale your way out of it if your trading expected value is negative which, if we don't mention the fraud for a second, seems to be what their biggest problem was.

blitzar 3y ago
> if your trading expected value is negative However, if you take the formula, assign yourself a positive trading edge and a large positive expected value you can justify placing whatever bet you want as a mathematical certainty. The most un-spoken about thing in all this is how exactly they went from +100bn to -8bn in a year.
lofatdairy 3y ago
I think that assumes that SBF and/or Ellison (I'm a bit less convinced Ellison thought they could trade their way out of the hole) saw negative expected value. I'm more inclined to believe that SBF, who's personal philosophy was a highschool-level understanding of Utilitarianism, completely misevaluated the EV Alameda + FTX's operation[^1], and that he took the wrong lesson away from JS -- that you can bet it all on something with +EV and if it doesn't work out, better luck next time. There's his interview with Tyler Cohen that's somewhat particularly telling[^2]: >And in fact it seems like Bankman-Fried, maybe more than anybody else I’ve ever heard of, really did think that way. In March 2022, Tyler Cowen asked him: >>COWEN: Should a Benthamite be risk-neutral with regard to social welfare? >>BANKMAN-FRIED: Yes, that I feel very strongly about. >>COWEN: Okay, but let’s say there’s a game: 51 percent, you double the Earth out somewhere else; 49 percent, it all disappears. Would you play that game? And would you keep on playing that, double or nothing? >And Bankman-Fried said yes! And Cowen went on: >>COWEN: Then you keep on playing the game. So, what’s the chance we’re left with anything? Don’t I just St. Petersburg paradox you into nonexistence? >>BANKMAN-FRIED: Well, not necessarily. Maybe you St. Petersburg paradox into an enormously valuable existence. That’s the other option. >That’s insane! You keep flipping a slightly weighted coin. Every time it comes up heads, the population of humans in the universe doubles. If it comes up tails once, every human in the universe is dead forever. How many times would you flip? Bankman-Fried’s answer was … as many times as it takes to kill everyone? Okay! Great! Nice utilitarianism you’ve got there. So, while I don't disagree that if you gave SBF or Ellison or anyone at Alameda a trade w/ obvious -EV that they'd take it. But I also think that what we see as clearly -EV was not being calculated as such at Alameda+FTX. I'm most inclined to believe that they saw it as either making trillions and saving the world or _not risking enough_ and not saving the world. [^1]: I don't know, maybe by factoring in a chance they solve AI extinction (lmao) by earning a trillion dollars and funding the right research groups, overestimating the chance that their bets pay off due to early success, or betting that using customer funds to fuel mass adoption which is key to making both operations profitable. [^2]: Quoted in Matt Levine's 10/5 newsletter: https://www.bloomberg.com/opinion/articles/2023-10-05/ftx-might-have-found-some-money https://www.bloomberg.com/opinion/articles/2023-10-05/ftx-mi...

blitzar 3y ago

> if your trading expected value is negative However, if you take the formula, assign yourself a positive trading edge and a large positive expected value you can justify placing whatever bet you want as a mathematical certainty. The most un-spoken about thing in all this is how exactly they went from +100bn to -8bn in a year.

lofatdairy 3y ago

I think that assumes that SBF and/or Ellison (I'm a bit less convinced Ellison thought they could trade their way out of the hole) saw negative expected value. I'm more inclined to believe that SBF, who's personal philosophy was a highschool-level understanding of Utilitarianism, completely misevaluated the EV Alameda + FTX's operation[^1], and that he took the wrong lesson away from JS -- that you can bet it all on something with +EV and if it doesn't work out, better luck next time. There's his interview with Tyler Cohen that's somewhat particularly telling[^2]: >And in fact it seems like Bankman-Fried, maybe more than anybody else I’ve ever heard of, really did think that way. In March 2022, Tyler Cowen asked him: >>COWEN: Should a Benthamite be risk-neutral with regard to social welfare? >>BANKMAN-FRIED: Yes, that I feel very strongly about. >>COWEN: Okay, but let’s say there’s a game: 51 percent, you double the Earth out somewhere else; 49 percent, it all disappears. Would you play that game? And would you keep on playing that, double or nothing? >And Bankman-Fried said yes! And Cowen went on: >>COWEN: Then you keep on playing the game. So, what’s the chance we’re left with anything? Don’t I just St. Petersburg paradox you into nonexistence? >>BANKMAN-FRIED: Well, not necessarily. Maybe you St. Petersburg paradox into an enormously valuable existence. That’s the other option. >That’s insane! You keep flipping a slightly weighted coin. Every time it comes up heads, the population of humans in the universe doubles. If it comes up tails once, every human in the universe is dead forever. How many times would you flip? Bankman-Fried’s answer was … as many times as it takes to kill everyone? Okay! Great! Nice utilitarianism you’ve got there. So, while I don't disagree that if you gave SBF or Ellison or anyone at Alameda a trade w/ obvious -EV that they'd take it. But I also think that what we see as clearly -EV was not being calculated as such at Alameda+FTX. I'm most inclined to believe that they saw it as either making trillions and saving the world or _not risking enough_ and not saving the world. [^1]: I don't know, maybe by factoring in a chance they solve AI extinction (lmao) by earning a trillion dollars and funding the right research groups, overestimating the chance that their bets pay off due to early success, or betting that using customer funds to fuel mass adoption which is key to making both operations profitable. [^2]: Quoted in Matt Levine's 10/5 newsletter: https://www.bloomberg.com/opinion/articles/2023-10-05/ftx-might-have-found-some-money https://www.bloomberg.com/opinion/articles/2023-10-05/ftx-mi...