tl dr ?
However, if a single person goes to the casino N times, their odds of going bust approach 100% as N gets larger, because as soon as you go bust you can no longer keep playing.
In any kind of iterated game, concern about tail risks is justified when the tail risk means you can no longer play the game.
Social science and economic research that attempts to analyze the "rationality" of human behavior has thus far failed to take this logic into consideration, falsely portraying people as irrationally "loss-averse". This oversight in the research community is obvious to successful investors, insurers, etc. because the latter have personal investment in the outcome of these iterated situations that the former simply do not. [This is only my attempt at a summary of his argument, not a defense]
[EDIT: spelling mistake]
And here’s where he goes off the tracks: even a quick search of Google Scholar will show that it, at best, grossly overstates the case to say that social science (and particularly economics, and even more specifically the subfield of “decision theory” that he calls out) research “thus far” has neglected tail risks in addressing rationality (it's also incorrect to say that there is some kind of consensus result that people vary from rationality consistently in the direction of loss-aversity.)
One of the reasons this issue keeps recurring is because, taken to its logical conclusion, the "time" probability paradigm (to use his term, there are better terms that have been used) as applied to persons' behaviors, only applies to that single individual, which then leads to a sort of paradox. Although we might want to know the probabilities for Theodorus Ibn Warqa, this is in principle unknowable because there is only one Theodorus Ibn Warqa. So when he is ruined, you can in theory say nothing about Maximillian Samuel Warqa.
At some point you have to invoke a counterfactual of sorts, and use a different person as a substitute, and this is why the "ensemble" probability model is invoked.
Taleb is right that the consequences of a ruin event are different at the individual level, but so what? The cost estimate is different from the probability estimate. It's not like people don't understand this. Maybe in the strawman literature he reads, but not in what anyone else is thinking. What's more typical is a ruin event one way or a ruin event another.
This is difficult stuff if you go beyond the toy gambler scenario he's dealing with. Just for one example:
https://dash.harvard.edu/handle/1/11718181
People wrestle with this not because they're idiots, but because it's more difficult than it seems once you really dive into it.
1. Taleb is incredibly smart 2. He is not being arrogant or trying to impress (any more than any other writer) 3. He is writing about something that is interesting and not everyone knows
I don’t see it how you do. That’s okay.
I’ve read a fair bit of his stuff and I like his approach to framing issues, which often ignores conventional wisdom, or phrasing.
The disagreement is whether he understands A. The issue and B. What others said. The whole of acadamia is full of different perspectives / why not here from Taleb?
Flatly stating that an entire field of inquiry has, up to the point at which you are publishing your commentary, failed to recognize and consider the ramifications of a phenomenon on which effect and it's ramifications numerous papers have been published in the field for several decades is not being aware of the factual circumstances and presenting them differently than others. Its either ignorance or dishonesty.
And when you compound that by using the mythical failure as a jumping off point for an explanation (“skin in the game”) for the systematic incompetence that seems necessary to explain how the field could fail to address the phenomenon, this compounds the problem.
1. Find some topic or idea that's interesting and fairly well-accepted but not receiving a lot of attention. 2. Claim ownership over the topic or idea as if the prior literature doesn't exist. Invent new terminology ideally; abandon existing terms to eliminate a trail to prior authors and ideas. 3. Present the idea or topic in a sensationalistic way. 4. When furor inevitably develops about lack of acknowledgement of previous writers and researchers and issues identified by them, frame argument as being about ad hominem critics versus solidity of your own arguments. Sidestep actual criticisms.
Well done.
It is actually well written, easy to read and briefly contains all the ideas in his later books.
His writing quality has decreased hugely since then, but his ego has not.