4 ms·

We expect some individual traders to beat the market (and some to do much worse than the marker); that's variance. But each individual trader should not expect

by jefftk 2y ago

We expect some individual traders to beat the market (and some to do much worse than the marker); that's variance. But each individual trader should not expect to beat the market, because they don't know if they're one of the lucky ones.

qwytw 2y ago
> But each individual trader should not expect to beat the market In aggregate sure. But unless we believe that it's entirely random some individual investors can still certainly expect to beat the market, they just can't verify that in advance.
- stouset 2y ago
  You’re being pedantic in all the wrong ways. I offer you a bet. We flip a perfectly fair coin. On every heads you gain 10% on top of your bet. On every tails you lose 10%. It is fair to say that after 100 flips you may profit. If one million people play this game, someone almost certainly will. But you can expect to lose money on this game. By the end, the average person will have about 60% of their original holdings (0.9^50 * 1.1^50). In this game it’s possible for winners to exist. It’s not even uncommon! You only have to get at least 53 out of 100 flips as heads. Unfortunately there’s also no function that lets you determine a winner in advance, and the longer you play this game the greater the expected loss. All of the available evidence shows that publicly-available actively-managed funds are essentially playing this game. As expected, many have incredible winning streaks… right until they don’t. Yes, Ren Tech’s Medallion Fund exists. But you can’t contribute to it; they don’t want your money. Because that requires scaling market inefficiencies and that in and of itself is an intractable problem. Novel strategies ripe for profit don’t have unlimited capacity. They rapidly exhaust alpha.
  - qwytw 2y ago
    > You’re being pedantic in all the wrong ways. No, I simply disagree with the whole premise, at least to a limited extent. > All of the available evidence shows that publicly-available actively-managed funds are essentially playing this game Yeah that's true, I was mostly talking about individual investors and/or non public funds.
- gbear605 2y ago
  That’s not what “expect” means in statistics. If we’re rolling 100-sided dice (each person rolls once), no person should expect to roll a 1, even though 1% of people will in practice roll a 1. Likewise, no one should expect to beat the market, even though many will in practice.
  - qwytw 2y ago
    > Likewise, no one should expect to beat the market, even though many will in practice. My only point is only that not every investor is rolling the same dice. It's just that it is effectively impossible to every verify whether you were rolling a 90-sided dice or a 100-sided one. It's rather clear that at least in the short to medium term (e.g 2-3 years) the stock is not even remotely perfectly efficient (that doesn't mean that the overwhelming majority of investors are somehow capable of utilizing that fact or that a significant proportion of those that did seemingly manage to do that weren't just lucky)
    - Dylan16807 2y ago
      If you don't know what dice you have, then the reasonable way to model that is a random choice of dice. And doing that gives you the same expectations as everyone using the same dice.

by jefftk 2y ago

qwytw 2y ago

> But each individual trader should not expect to beat the market In aggregate sure. But unless we believe that it's entirely random some individual investors can still certainly expect to beat the market, they just can't verify that in advance.

stouset 2y ago
You’re being pedantic in all the wrong ways. I offer you a bet. We flip a perfectly fair coin. On every heads you gain 10% on top of your bet. On every tails you lose 10%. It is fair to say that after 100 flips you may profit. If one million people play this game, someone almost certainly will. But you can expect to lose money on this game. By the end, the average person will have about 60% of their original holdings (0.9^50 * 1.1^50). In this game it’s possible for winners to exist. It’s not even uncommon! You only have to get at least 53 out of 100 flips as heads. Unfortunately there’s also no function that lets you determine a winner in advance, and the longer you play this game the greater the expected loss. All of the available evidence shows that publicly-available actively-managed funds are essentially playing this game. As expected, many have incredible winning streaks… right until they don’t. Yes, Ren Tech’s Medallion Fund exists. But you can’t contribute to it; they don’t want your money. Because that requires scaling market inefficiencies and that in and of itself is an intractable problem. Novel strategies ripe for profit don’t have unlimited capacity. They rapidly exhaust alpha.
- qwytw 2y ago
  > You’re being pedantic in all the wrong ways. No, I simply disagree with the whole premise, at least to a limited extent. > All of the available evidence shows that publicly-available actively-managed funds are essentially playing this game Yeah that's true, I was mostly talking about individual investors and/or non public funds.
gbear605 2y ago
That’s not what “expect” means in statistics. If we’re rolling 100-sided dice (each person rolls once), no person should expect to roll a 1, even though 1% of people will in practice roll a 1. Likewise, no one should expect to beat the market, even though many will in practice.
- qwytw 2y ago
  > Likewise, no one should expect to beat the market, even though many will in practice. My only point is only that not every investor is rolling the same dice. It's just that it is effectively impossible to every verify whether you were rolling a 90-sided dice or a 100-sided one. It's rather clear that at least in the short to medium term (e.g 2-3 years) the stock is not even remotely perfectly efficient (that doesn't mean that the overwhelming majority of investors are somehow capable of utilizing that fact or that a significant proportion of those that did seemingly manage to do that weren't just lucky)
  - Dylan16807 2y ago
    If you don't know what dice you have, then the reasonable way to model that is a random choice of dice. And doing that gives you the same expectations as everyone using the same dice.