Hacker News Reader

Comment by croes | Hacker News Reader

Ever heard of survivorship bias?

Yeah but why do those who survive survive?

You know that old scammer trick? Pick 512 wealthy people, pick a stock, write half of the the people the stock will rise, the other half the stock will fall. Wait for the real development, dump the addresses of the losers, pick a new stock and start again. Write one half stock will rise, other half, stock will fall. Repeat until only one address is left. Now give him the opportunity of an large investment. You have shown your skills because you were right 10 times in a row.

And how does this apply to modern investment funds who have to regularly provide client reports and make disclosures?

Take 512 investment funds, in one year half of them make a profit, half of them will make losses. Drop the ones that had losses, the next year half will make profits, the other half will suffer losses. Repeat until one fund is left: "Invest in us - we are market experts who had no losses for the past 9 consecutive years!"

Well yes, I understand that. The point is that your model predicts the current landscape of funds as being about as likely as pigs flying.

You can invest in Apple stock in the early 90s because you are a savvy investor who sees the massive potential, or you can invest in them because the stock shows up at the top of an alphabetical list. Either way you're rich, and get called a brilliant investor, while all the people who invested in ZZCO because they picked the bottom of the alphabetical list, or some other similarly random choice, are bankrupt.

Of course you can invest at random and it will sometimes work. But it does not explain the realities of the industry. Even just one of the top funds has an infinitesimally small chance of existing randomly, let alone a glut of them.

The impeccable track record of index funds suggests that "Even just one of the top funds has an infinitesimally small chance of existing randomly" is not actually correct.

Well index funds aren't 'the top funds', they're the benchmark against which the alpha of the top funds is measured. If we change the framing to over/underperforming the S&P500 I don't see how the argument changes. RenTec is up 66% per year since 1988, how could that possibly be random?

Because people don't know the performance of dart throwing monkeys https://www.forbes.com/sites/rickferri/2012/12/20/any-monkey...

That article explains that the outperformance was due to a broader economic trend of small cap companies outperforming weighted indices. They effectively baked in a known-to-win strategy from the start and then ran it in various slightly-randomised ways (and they're fully open about it!)

One big early hit can keep paying off. And some funds have long lock-ins.

In this case, they sold part of their tencent to buy this. Selling a little bit each year of a superstar early investment shows good earnings each year.

I don’t know this fund at all, so they could be awesome for all I know. But saying “they are super smart for buying tencent” is not useful without lots of other info.

Usually people who say that without extra evidence or mentioning survivor bias aren’t properly aware of what’s what.

Mostly it’s random chance. The reality of investing is that it’s mostly down to chance. Take an equally skilled set of investors and give them the same market access and some will do exceptionally well.

This does not stand up to scrutiny at all, for any kind of investing. There have not been so many VC/hedge/PE funds in history that the top ones would arise through chance.

I mean you can get ten thousand monkeys throwing darts at a board full of companies as an investment strategy, and a handful of those monkeys will be top performers with excellent returns.

How many vc/hedge/pe funds were there? How many funds around today were here 20 years ago.

Failures don’t typically advertise and it’s quite likely there ARE/WERE thousands or millions of small firms managing a few million each.

Simple answer, it's impossible that all people choose the wrong investment even if you just choose by chance. It like playing the lottery, why do some people win? Because not all can lose all the time. And as soon you reach a certain wealth level it's easier to keep that level than to reach it in the first place, unless you are in idiot.

How many times can one person win the lottery? The large firms seem to successfully bet on winners quite a bit.

_Somebody_ wins it every week.

People lose money in lottery tickets, casinos, equity investments, etc all the time. If a VC is successful over a period of time with a history of investing in winners, then there is something more than simply randomness.

With enough VC firms, and enough small bets, that's not guaranteed to be true.

Luck, usually.

Mostly luck.

To quote Mazarin, "One must not ask of a general, 'Is he skillful?', but rather, 'Is he lucky?'". I think what Mazarin was probably getting at was that the best generals will always have others claiming that their success was due to luck.

I hate when people bring up survivorship bias about something like this. It's such a negative outlook. It's like a "don't even try because it's just a fluke that happened to them" type of mindset. It doesn't help anyone.

I read this more as “make sure to account for survivor bias and don’t say stupid stuff.”

Survivor bias doesn’t mean that everything is a fluke. It just means that some things are flukes and it’s important to include that in our models so we don’t think something random is non-random.

This is very different than everything being 100% random.

Reply on news.ycombinator.com