5 ms·

It's a nice piece of math, but it just confirms common sense. If you have a small edge and the payoff is very likely, you bet a decent chunk of cash. If you ha

It's a nice piece of math, but it just confirms common sense. If you have a small edge and the payoff is very likely, you bet a decent chunk of cash. If you have a big edge and payoff is unlikely, you bet a half decent chunk of cash. If you have a big edge and the payoff is very likely, you bet a double decent sized of cash. If you have a small edge and payoff is unlikely, you bet a little bit. If you don't have an edge, you don't bet.

ultra_nick 5y ago
You're describing the anti-martingale which is a different strategy with a statistically expected profit of $0 for random coin flip type games. The Kelly criterion is different and has a positive expected profit.
- danielmarkbruce 5y ago
  No, that's not what I'm describing at all. Re-read it, or re-read what martingale/anti-martingale is. I don't say anything about betting size based on a previous bet.
modeless 5y ago
You'd think the optimal strategy is common sense but for small probabilities and large payoffs it really isn't. Many people simply ignore very risky investments on the theory that they're just too risky, when the actual optimal thing is to invest small amounts if the payoff is big enough.
- danielmarkbruce 5y ago
  Do people really ignore this? The entire VC industry seems dedicated to these exact bets. The guys at Sequoia for example are approximating the kelly criterion and likely don't ever think about it explicitly.
SkyMarshal 5y ago
> If you have a big edge and payoff is unlikely Why would the payoff be unlikely if you have a big edge?
- CobrastanJorji 5y ago
  Say there's a horse race. Paul Revere is listed as having 500:1 odds of victory. You, however, know that because the weather is clear, he actually has a 5% chance to win. It's a great idea to bet on Paul Revere, but you're still most likely going to lose your money.
  - danielmarkbruce 5y ago
    Exactly. And common sense tells you a) you want to make the bet, but b) that you don't want to bet 50% of your bankroll on this. Maybe 5% makes sense.

ultra_nick 5y ago

You're describing the anti-martingale which is a different strategy with a statistically expected profit of $0 for random coin flip type games. The Kelly criterion is different and has a positive expected profit.

danielmarkbruce 5y ago
No, that's not what I'm describing at all. Re-read it, or re-read what martingale/anti-martingale is. I don't say anything about betting size based on a previous bet.

modeless 5y ago

You'd think the optimal strategy is common sense but for small probabilities and large payoffs it really isn't. Many people simply ignore very risky investments on the theory that they're just too risky, when the actual optimal thing is to invest small amounts if the payoff is big enough.

danielmarkbruce 5y ago
Do people really ignore this? The entire VC industry seems dedicated to these exact bets. The guys at Sequoia for example are approximating the kelly criterion and likely don't ever think about it explicitly.

SkyMarshal 5y ago

> If you have a big edge and payoff is unlikely Why would the payoff be unlikely if you have a big edge?

CobrastanJorji 5y ago
Say there's a horse race. Paul Revere is listed as having 500:1 odds of victory. You, however, know that because the weather is clear, he actually has a 5% chance to win. It's a great idea to bet on Paul Revere, but you're still most likely going to lose your money.
- danielmarkbruce 5y ago
  Exactly. And common sense tells you a) you want to make the bet, but b) that you don't want to bet 50% of your bankroll on this. Maybe 5% makes sense.