[vulcan01]

Why would they be fired, if they were one of the top 50 people in the bank? From a pure profit perspective, if this guy was that good, did it really matter if he was using 'sophisticated algorithms' or not?

(I may have misunderstood your first comment. If so, please let me know.)

[__alexs]

The risk profile of someone who is consistently lucky is probably quite different from someone who is consistently right.

[PakG1]

The interesting question is whether he developed any tacit skill or tacit knowledge that would explain his consistent success but would also be difficult to communicate or explain precisely because it's tacit, and therefore can only be described as listening to his gut. In the article, something like this is described:

I can tell she is going to say something else, and I’m pretty sure I know what it is. She’s going to share with me how much time she has left. I can hear it in her pauses. After so many years working the phone, I’ve learned to pick out the nuances, the things being said behind what’s being said, entire life stories even, in a hesitation or vocal inflection, in blank moments in time.

It's difficult to explain what makes a pause have meaning or nuance, and yet, this guy may be able to interpret them consistently in a simple phone call without seeing facial expressions and body language. Tacit stuff is really tough to define and nail down. So it's easy to say gut feeling instead. But as tacit stuff finally gets clarified, defined, and explained, we're often able to see that tacit stuff is anything but a gut feeling.

[bitL]

College education sometimes kills some major talents so one could also imagine that people who avoided higher education might keep some rare talents distinguishing their performance from the others. Tiny chance but it might happen.

[DANmode]

The entire metric is "beating the market".

If someone showed me airtight, verifiably top tier calls over 15 years (especially the last 15) I'd probably rub whatever rabbit's foot they asked.

[nullc]

Are you aware of the old random stock pick fax scam?

You get some huge list of fax numbers. You pick some random penny stocks and fax each one a different random pick, telling them it's going to explode. Then you look and see which picks fail and you discard those numbers.

Eventually you get down to a small list of numbers they you've been consistently making amazingly prescient stock picks to for the last weeks/months... then you tell them all to invest in some dogcrap that you've cornered or ask them to buy your latest picks, or some other story to extract a lot of money from them.

The victims rubs your rabbit's foot and you cash in.

Thing is that the same scam can be performed in a decentralized manner, even by accident, just by having a lot of investment advisors some of whom get a consistently lucky run.

[DANmode]

15 years is statistically unlikely to be a lucky run, is the point unaddressed.

[nullc]

In the space of all fund managers / investment advisors? I think not at all unlikely.

[__alexs]

Say you have a room full of millions of people playing roulette. You have no knowledge of how they place bets but you know how roulette works. There's a wheel, 38 slots etc.

You watch all the players for a while and see that 1 player has won every single bet they've made 10 times in a row.

How much money would you be willing to give that person to bet for you?

Now imagine that you talk to this player and they reveal to you that actually they work for the company that made the roulette wheels and that they can control where the ball lands 99% of the time with the computer hidden in his shoe. They even show you the computer and do some example bets to prove it.

How much money would you give them now?

[fiddlerwoaroof]

I’m not sure you can dismiss someone being “consistently lucky” so easily.

https://en.m.wikipedia.org/wiki/Superforecaster

The thing that would concern me more is the pattern of dishonesty than that the person did well over a long time by just going with their gut.

[fauxpause_]

> Superforecasters do not predict the future with perfect accuracy: Bloomberg notes that they made a prediction of 23% for a leave vote in the month of the June 2016 Brexit referendum. On the other hand, the BBC notes that they accurately predicted Donald Trump's success in the 2016 Republican Party primaries.[11]

Is there any good documentation on superforecasters efficacy? Like with actual statistical measurements? Because the Wikipedia article is really dumb.

[getlawgdon]

The longer the horizon, the less so.

[karaterobot]

I assumed they meant top 50 on the org chart, not necessarily the top 50 in objective performance over a long term. I could be wrong.