undefined | Better HN

0 pointslazide5y ago0 comments

It's also really hard to pick 'can beat the market' (predictive of future behavior), from 'have beat the market' (retrospective analysis of results - what all it's associated survivorship biases and bias towards success in a past which will never repeat itself exactly.

Someone WILL always beat the market. A small set of people WILL beat the market consistently - until they don't. That set gets smaller every year, until over a long enough timeframe it goes to zero.

History has shown predicting who will be able to win over a specific timeframe is essentially impossible, and more random chance than skill - with temporary counter examples that always revert to the mean over time.

0 comments

3 comments · 1 top-level

fractionalhare5y ago· 2 in thread

History has not shown that. It is easy to say that people beat the market by chance, but you should actually try calculating how many standard deviations from the mean some of these track records are. There haven't been anywhere nearly enough hedge funds to justify the most consistent track records being due to chance alone.

For illustration: assume any given fund has returns which simply approximate a normal distribution; i.e. their returns are theoretically just noise. Then the chance of the fund achieving a 2 sigma return in any given year is about 2%. We can model the odds of such a firm consistently exhibiting a 2 sigma return for 20 years in a row using a binomial distribution with n = 20 trials, k = 20 successes and success probability p = 0.02. Then we have

binom(20, 20) * 0.02^20 * 0.8^0 = 1x10^-34

There are firms which have consistently beaten the market by a significant margin for that long. Even if you relax the constraint to 10 years, you still get "only" 1x10^-17. At a certain point this becomes similar to saying that Steph Curry isn't actually good at basketball, all of his 3 point throws are just the expected outcome of lots of mediocre players existing who didn't make it to the NBA.

I do agree that retail investors should just invest in index funds though. And I agree it's extremely difficult to determine who has the genuine skill to beat the market before they've beaten it for so long that they're no longer accepting money.

lazideOP5y ago

If you include their costs and fees, that is not the case though is it? Or at least the data I've been able to find doesn't support that.

You're also talking about rearward looking data, not forward looking predictability, often only including hedge funds for instance that still exist today. For any given hedgefund who outperformed in the past, how likely is it that, if you made a judgement today on their ability to perform, and then waited 1, 5, 10 years in the future, it would match that judgement? What are the odds they'll exist then, or be merged or renamed or whatever?

If you today made a bet for 10 years from now as to Steph Curry's 3 point average then, how likely would it be to win?

There is a reason anything to do with the markets is almost always followed with 'past performance is no guarantee of future results'.

There are also confounding effects like the Bernie Madoff's and Enron's of the world, and major issues with visibility into most hedge funds that make it difficult to actually compared performance on a level playing field. If someone is 'loading the die' using insider trading, tips from the CIA or whatever, it would be really difficult to call them anything but incredibly skilled/lucky - and you couldn't tell the difference unless it blew up publicly.

fractionalhare5y ago

Like I said, I agree that it's extremely hard to judge if someone has genuine ability before they exhibit the track record. I'm simply saying that once the track record is established, attributing it to chance alone doesn't really make sense mathematically speaking.

1 more reply

j / k navigate · click thread line to collapse