User Tools

Site Tools


The Black Swan

The Impact of the Highly Improbable
Nassim Nicholas Taleb

failed to fetch data: unkown error

Confirmation bias: we seek information that confirms our theories rather than actually testing them by seeking to refute them.

The problem of induction. He seems to go so far as saying that finding evidence consistent with the theory is meaningless, which I don't think I go that far. It is true that you can find out the limits of the theory better by trying to falsify it, but it is surely better than nothing to just test the theory against any evidence you happen to stumble across.

I think that confirmation bias may be explained at least in part by the rationalizing brain. We've come up with a theory, so we want to make a story about why it's a true theory (see narrative.)

Round-trip error: confusing “no evidence for X” with “evidence for no X”. From a logical perspective, people tend to confuse implication with equivalence. Most terrorists are moslems, so most moslems are terrorists.

Here, and in general, he argues that such failures of logic are evolutionarily tolerated because we evolved in a simpler environment where such fine distinctions where not so important. He says that confusing “most killers are wild animals” with “most wild animals are killers” was not so problematic. I'm not convinced. Is that really so harmless? Why are we so sure these errors are very bad now? Maybe the errors are tolerated because they are not so bad and there is no practical alternative (this is certainly the case with narrative.)

Not applying knowledge in different contexts. Statisticians act statistically naive outside of their professional context.

We have specific inductive instincts. Children know to infer that when presented with a picture of a person that the person's group is likely to have the same skin color, but not to be fat just because the one person is fat.

Tunneling: focus on a few uncertainties

Narrative fallacy

Our vulnerability to overinterpretation and out predilection for compact stories over raw truths.

This is where he touches on the rationalizing behavior, but I think his spin is not quite right. For example, in talking about the decision literature, he identifies two systems. In system 1, the intuitive system, we are quick, but make all these errors. System 2 is the proper rigorous rational thought which most people don't do. The thing is, he identifies narrative as a cause of unsound system 1 thinking, whereas it is more an effect. We have used unconscious judgment to come up with a viewpoint, then we seek a story to explain it.

People are persuaded by stories and anecdotes and not by solid quantitative evidence.

Narrative and therapy: journalling

Keeping a journal makes you feel better, he says because after narratizing your life it acquires an air of inevitability, and so your personal responsibility is diffused (or perhaps the responsibility of others too.)

Well, I think this may be true to some degree, though I have thought of it as a way of developing understanding of what has happened.

It seems that in general in his zeal to puncture spurious explanation (e.g. in journalism the habitual offering of explanation for market moves), he slights the value of narrative or theories. Of course, narrative, stories, examples, theories are essential for persuading others of your ideas, but also good theories can be useful even if you keep them to yourselves.

Dopamine and pattern detection

Perception of pattern is biologically regulated, and can be disinhibited resulting in the spurious perception of patter, e.g. with the L-dopa takers who become compulsive gamblers because they see patterns in randomness.

It is also interesting the context of me personally seeing all these patterns when I was elated a couple of weeks ago, and of the grandiousity of manic episodes.

Black swan hunters * respect v.s. money * doesn't make you happy ? * find a group who agree with your choice for the unconventional.

“Living in the antechamber of hope”

He's sort of puzzled by the artists, scientists, etc., who labor without recognition, compensation or status in the hopes of hitting the big time. He sees this as a positive black swan. You gamble on striking it rich with your life. He doesn't think it makes you happy because he says that even rewards are what make you happy.

He says may be explained by Aristotle's idea that seeking virtue rather than seeking pleasure makes you happier.

He hasn't realized the power of the idea that we may feel compelled to behave in ways that don't make us happy.

This topic is clearly in the same area as my discussion of status and meme propagation strategies.

Respect for elders: Elders know black swans

Silent evidence

The anthropic principle

Evolved optimism: being optimistic is adaptive, and this is why people disregard the evidence of bad outcomes for those who are in the graveyard.

This is somewhat related to my idea of the sessile strategy, but from the opposing direction. It is true that by merit of surviving we may take a spuriously rosy view, but there is information in the fact that we have survived. It is true that luck may have played a big role, but who knows, and anyway luck is not controllable so may not matter that much that we neglect it. Also, more along the lines of evolved optimism, maybe counting on getting lucky is a viable strategy. But back to the sessile strategy. It's a reasonable guess to keep doing what we've been doing because it has worked so far. In particular, the sessile strategy is attractive for organisms that are overmatched by their threats, e.g. limpets v.s. seagulls, where we are not equipped to perceive or process the threat. But it seems that a lot of what Taleb is saying is that we are overmatched by our threats, so sticking with what has worked is perhaps adaptive.

In fact, in the next section where he argues about the futility of prediction he says that prediction experts don't do any better of than the simple rule that tomorrow will be like today. This is the rationalized sessile strategy in a nutshell. Implicitly, any expert that seems to be succeeding at prediction is either using the sessile strategy, was just lucky, or is good at post-hoc rationalization.

How about the psychological immune system?

Kahneman and Tversky Judgment and decision making

What I learned losing a million dollars

Cognitive diversity

The long tail

Robert Trivers

Dan Sperber, contagious mental categories (memes)

“But I find the emphasis on economic inequality, at the expense of other types of inequality, extremely bothersome. Fairness is not exclusively an economic matter; it becomes less and less so when we are satisfying out basic material needs. It is the pecking order that matters!” p 237.

Religion ended the reproductive domination by the alpha male. Could it solve the problem of intellectual dominance?

Taleb mentions flow as supporting the idea of the desirability of even rewards, thus reinforcing the mystery of the swan hunters. But anyone who has done it knows that swan hunting leads to flow, and is thus pleasurable.

books/black_swan.txt · Last modified: 2010/05/24 13:51 by ram