[dsacco]

There is survivorship bias, but the coin flipping analogy is a tired trope (and it implies something utterly different). There aren't a million traders, and there are more than 20 consistently beating the market. Significantly more. Do the actual math with some attempt at empirical numbers and your argument might make sense. Until people start doing that, repeating this line about coin flips and the law of large numbers doesn't add anything to the discussion.

First of all, you haven't qualified who you're including in the set of "traders" and who you're including in the set of "winners." Is everyone who signs up for ETrade included? That's like major league baseball players being judged the same as way as high school baseball players. More importantly, have you done the cursory research to account for funds that consistently beat the market? How do you account for the firms that beat the market over periods that span decades? Just ridiculously lucky? What counts as a coin flip? A single trade? A trading day? Are the coins summed per trader or per fund? How are we quantifying this assessment?

It's like every time someone brings up the coin flipping analogy they use these outrageous numbers without any attempt at citing a grounded source in reality. As I say every single time in threads like this: yes, it's incredibly difficult to purposefully and consistently beat the market, but that's worlds away from impossible. There is information asymmetry in the market, relatively few people/firms are capable of identifying alpha based on that, and fewer still are capable of capitalizing on it. But they exist!

Stubbornly repeating the coin analogy is like insisting on proof that basketball players have inherent skill instead of luck, because most people can't make it to the NBA. We have clear examples of firms beating the market consistently for decades at a time, net of fees. I am personally familiar with people whose strategies profitably trade on small pockets of predictable events in timeseries tick data. Their strategies are smaller (~high 6 - low 7 digit accounts using personal capital), but they consistently earn 27-30% each year by trading strategies that are too capacity-constrained for larger firms (and usually they do this after being in the industry for some time).

I make this point not to pick on you (it's not personal!), it's just that I see this repeated in every thread related to trading on HN. Referring to trading as coin flipping when your familiarity mostly stems from news reports flies in the face of people who are capable of developing profitable trading algorithms and who have seen it. It's as if someone told you that it's impossible to develop well-engineered software. It's exhausting. There's this weird leap from (correctly) concluding that most people in the industry can't beat the market, to damning the entire concept.

If you want to say there is a lot of survivorship bias in the industry, sure, I'll agree with that. But what's the point of using Fama's coin flipping analogy if there's no rigor behind it? It precludes so much nuance in fund performance. Many funds can't beat the market at all. Many do beat the market, but they purposely decide to eat away all those gains with fees when they could run a far leaner ship. And the elite do consistently beat the market, until they eventually get large enough to diversify into multiple funds (accepting that most will be mediocre) or they return investor money because they don't need it and their strategies are capacity constrained.

[soVeryTired]

Ok, fair enough. The coin flipping analogy isn't great. But it's a pithy way to make a very valid point: there's a massive amount of luck in investment. To see this, let's think about coin flipping again :).

Suppose I have a 90% chance of making a positive return in any given year. The chance that I'll make money every year for ten years is about 35%. Over 20 years, it's about 12%.

Now, I don't believe that anyone has a 90% chance of making money in a given year. Even if I conceeded that it's possible to have a long-term edge in the markets, I'd say it's more like a 55% chance of making money. So instead of there being a handful of elite geniuses with a 90% edge, I think it's more likely that there are many smart people with a 55% edge, all tossing coins.

I've hidden a lot of things away in this analysis, and I'd need to write a full essay to give the issue its due. For example, I've ignored the possibility to "beat" the market by leveraging up your S&P exposure and charging fees on top. If the S&P goes up, you beat it. If the S&P goes down, you blow up and start again.

[FabHK]

What you say might be right. But there is the purely ontological question (do traders exist that outperform consistently?), and the different pragmatic question: Does it make sense to part with my money and give it to active asset managers?

The answer to the theoretical first question might well be "yes, there are exceptional traders outperforming the market". The second "real-world" question is much more involved:

* can I reliably identify them, ex ante? That's pretty hard.

* do I have access to them? Many of the examples cited (Renaissance Tec, individual traders trading personal account) are closed to outside investors. Many golden investment opportunities are channeled to people that are very well connected (either in the finance industry, or old money, or close to politics). Similarly, there might be legal access restrictions - you must be a qualified investor, or reside in a certain jurisdiction, etc. etc.

Personally, I think most of the asset management industry is basically a giant rent seeking exercise, charging extraordinary fees (over the long run) for very little value. There are exceptional performers, but they are not accessible to most people. (As a side note, much of their exceptional performance might come from insider information more than exceptional analysis.)

Thus, the standard advice stands: put most in cheap index funds/ETFs, and maybe develop some expertise in certain areas and dabble in it with a fraction of your assets, if you're so inclined.

The reasons that the finance industry is particularly prone to unproductive money skimming (rent seeking) include:

* massive information asymmetry (similar to real estate)

* psychological biases (most people don't realise that someone taking a tiny 2% fee per annum has basically taken half your savings by the time you retire)

* mostly experience goods (or even post-experience goods): enormous difficulty of evaluating quality ex ante, or sometimes even ex post (your pension pays you X per month. Could a competitor have done better?)

* regulatory capture

etc.

The coin analogy is a perfectly legitimate parable to highlight how difficult it is to evaluate asset manager performance. Sure, the actual analysis then still needs to be done, but the outcome, from what I can tell, remains pretty damning.

[lordnacho]

> The answer to the theoretical first question might well be "yes, there are exceptional traders outperforming the market". The second "real-world" question is much more involved: > * can I reliably identify them, ex ante? That's pretty hard.

The answer to question one, as you say, is yes. If it was purely coin tossing any group of performers would get halved each year (let's not mention skew). Now there are funds that fall quite hard, like Odey. But it's clear there are funds that seem to defy gravity.

Question Two. Here's how I decide if I believe in a fund.

Here's a real life example. A friend of a friend came to a meeting and explained that under certain circumstances, you can buy certain unit trusts cheaper than they're worth. You need to dig out the nitty gritty details of each fund: the holdings, the rules, the fees, trading regulations, and so on. You put all this in a computer, which tells you when there's a mispricing. You need good relations with various counterparties, or you won't get the trade. And you can only do it up to a certain scale, because there's only so much mispricing. And it's sensitive to costs, so you need his particular cost agreements.

This also explains why you'd have to invest in this particular guy. He has the infrastructure and relationships already in place, so even if you knew his method, you couldn't do it yourself.

You're right about a lot of funds being smoke and mirrors though. I used to run one, and the investors never ask the right questions. Mostly they chase returns, but they don't even know how to tell the difference between one track record and another. It makes a huge difference how they're generated, yet most questions are simply trying to put you in a category, like produce in a supermarket.

[chollida1]

> You're right about a lot of funds being smoke and mirrors though. I used to run one, and the investors never ask the right questions.

Wow!!!!!

I've never heard an owner of a hedge fund refer their fund as "smoke and mirrors". I really hope for your sake that you just misspoke.

I'm sorry it didn't work out for you. I understand how hard the industry can be, if you ever think of trying again, don't let your past failures dissuade you.

I'm available to chat if you'd like:)

[lordnacho]

We've already spoken, well emailed. I'm doing prop for an HFT, not a million miles from running a fund. Well, it's better anyway.

As for smoke and mirrors, naturally I mean other people's funds! Mine actually would have won the bet with Warren by a fair margin.

Come to think of it, I was a partner before that in a fund that wasn't so good, but had lots and lots of investors climbing over each other to invest.

[FabHK]

> the investors never ask the right questions. ...

> most questions are simply trying to put you in a category, like produce in a supermarket.

Yeah, and who can blame them? It's very hard unless you have an inside edge. And lamentably many funds cater to those investors, put out nice and shiny brochures, and get good amounts of money to manage.

[gmarx]

The coin flip being tired or a trope is irrelevant. The million and 20 may not even be correct (haven't checked). the point is that people see patterns even where they don't exist. That's why science has methods for telling whether a pattern really represents something or if it would likely be seen even if the data were random. Until you do those stats on hedgefund returns your irritation at the trope remains unconvincing.

[dsacco]

I suppose we're at an impasse then, because the claim wasn't mine; nor is it even sufficiently formalized as to be falsifiable (which really was my point here). We cannot apply scientific methods to claims that are only semantically meaningful. For example, the definition of "coin flip" was never even specified, which is why the constant use of the analogy is absurd.

When firms like RenTec exist and continue to empirically generate market-beating returns over 20-30 year timespans, the burden ceases to be on the critic of a claim to empirically disprove it, especially if it's not even falsifiable. Here, you are doing the same thing as the previous commenter, except you're not using the analogy.

You cannot open your argument with the premise that returns are purely stochastic if that's not self-evident - you need to prove that. But I have never seen a single individual attempt to quantify the analogy, not even in a forced way to make it support their thesis. It's taken for granted that superlative returns are purely chance, and the goalposts are constantly moved whenever someone brings up successful funds.

[gmarx]

I understand you have declared we are at an impasse so I will just take the "last word". Your second paragraph is mere assertion. I assert the opposite. The existence of extreme outliers is exactly what we see in statistical distributions all the time. Being able to identify them in retrospect is childs play.Doing it in advance is the trick. That fund very likely does have great skills and plans at work that happen to have been met with circumstances that made them work. Yes you can attribute that to skill.

[z0r]

All posters are more or less right in this comment chain, but talking about coin flips and renaissance in the same breathe _seems_ to be silly, if you've read about the latter. (Not an expert so I might have been mislead by what I've read...)

[adrianN]

> it's impossible to develop well-engineered software

That doesn't sound so stupid.

[HeyLaughingBoy]

It does to people who've done it consistently.

[MockObject]

And if it were really like coin flipping, wouldn't the average industry profit be zero?

[fosco]

https://xkcd.com/1827/