139 lines
4.9 KiB
Markdown
139 lines
4.9 KiB
Markdown
---
|
|
title: _Fooled by Randomness_
|
|
tags:
|
|
- status/complete
|
|
- topic/risk
|
|
- type/media-commentary
|
|
---
|
|
# _Fooled by Randomness_
|
|
|
|
This is the commentary companion to [[taleb_2001_fooled-by-randomness]].
|
|
|
|
## Review
|
|
|
|
This is a low-effort bestseller bait
|
|
on the order of the lucky fools it (rightly) criticizes.
|
|
The only difference is that Warren Buffet is a successful trader.
|
|
|
|
If you've read or are reading this book
|
|
and find yourself agreeing with any of its conclusions,
|
|
try _The Failure of Risk Management_ by Douglas W. Hubbard.
|
|
Hubbard dismantles Taleb's pessimism point by point,
|
|
and it's an excellent read besides.
|
|
|
|
One might be tempted to say
|
|
that the damage Taleb and this book have done
|
|
to risk management and adjacent fields is immeasurable,
|
|
but I'm confident a rigorous model could estimate the damage
|
|
to an acceptable level of certainty.
|
|
|
|
## Critiques
|
|
|
|
### Intellectual Insecurity
|
|
|
|
So far I've gotten the distinct impression
|
|
that statistics hurt Taleb's head,
|
|
and he is intimidated by academics.
|
|
|
|
#### "Logic Without Statistics"
|
|
|
|
FbR uses few citations,
|
|
relying on the strength of Taleb's logic alone, by his own stating.
|
|
|
|
Taleb argues this strategy is perfectly legitimate, which is _true_,
|
|
but it does not follow that it makes for the most robust argument.
|
|
In recent editions of the text, Taleb claims that his editors have _implored_ him
|
|
to provide figures, graphs, studies, etc. as---_he agrees_---would be expected
|
|
for any similar book on statistical phenomena, but he refuses.
|
|
|
|
I don't find this approach charming at all,
|
|
especially considering how critical Taleb is of demagogues like Warren Buffet,
|
|
who could write that their success was foretold burning oracle bones
|
|
and the book would still be a bestseller.
|
|
_Fooled by Randomness_ fails to differentiate itself from such books,
|
|
except in that it was written by a mediocre trader.
|
|
|
|
Juxtaposing FbR with [[hubbard_2020_failure]],
|
|
which is a more traditional work of statistical thought---
|
|
well researched, and with a thorough bibliography---
|
|
Taleb's arguments are considerably less satisfying.
|
|
When Hubbard is wrong, it's clear his interpretation is flawed in that instance,
|
|
when _Taleb_ is wrong, I question the foundation of all his arguments.
|
|
Less politely, I wonder why I'm listening to him just make up justifications
|
|
for what he already believed.
|
|
|
|
So far I'm lead to believe what Taleb means by "logic"
|
|
is only anecdotes and aphorisms.
|
|
He wants to be Plato, but he comes off as "this came to me in a dream".
|
|
|
|
> [!quote] Chapter 2? (pp.)
|
|
> Scientists can not meaningfully describe the probability of black swans
|
|
> because it would require observing the future.
|
|
|
|
> [!quote] Chapter 2? (pp.)
|
|
> Accountants don't care about probability,
|
|
> if they did they wouldn't be accountants,
|
|
> and if they were they would make an error on your tax returns.
|
|
|
|
#### Qualitative Probability
|
|
|
|
Taleb loses me in the introduction
|
|
when he states that he defines _probability_ qualitatively.
|
|
|
|
[[hubbard_2020_failure]] gives a comprehensive history
|
|
of the terms **uncertainty**, **probability**, and **risk**,
|
|
|
|
Later it's clear he what he means by probability is **uncertainty**.
|
|
|
|
> [!quote] Chapter 2? (pp.)
|
|
> Certainty is something likely to occur in the largest number of possible worlds,
|
|
> uncertainty concerns what is unlikely to occur in many possible worlds.
|
|
|
|
#### The Black Swan
|
|
|
|
**The black swan**, or the unforeseen event,
|
|
is the idea that no quantitative risk management is possible
|
|
because of the possibility of a single loss
|
|
that would negate all previous gains.
|
|
|
|
Hubbard points out that Taleb's position is paradoxical.
|
|
|
|
> [!quote] [[hubbard_2020_failure#A Note About Black Swans]]
|
|
> ...he is assessing the validity of using historical examples
|
|
> by using _historical examples_.
|
|
|
|
Besides its credibility, the suggestion reeks of intellectual insecurity.
|
|
It is very convenient to dismiss the whole of statistics based on logic alone,
|
|
much more so to dismiss the tests used to prove the validity of statistical methods.
|
|
|
|
### False Lucky Fools
|
|
|
|
Taleb repeatedly conflates legitimate [[lucky-fools|lucky fools]]
|
|
and people with ideas he doesn't like.
|
|
|
|
> [[hubbard_2020_failure]]
|
|
> does a much better job of explaining "lucky fools"
|
|
> (see [[the-failure-of-risk-management#Red Baron Effect]]).
|
|
|
|
#### Nero Tulip vs. John
|
|
|
|
Taleb uses the story of Nero Tulip,
|
|
an incredibly cautious investor,
|
|
and his success over his risk-loving rival "John"
|
|
to promote the idea that modern quantitative methods
|
|
(as John is presumed to represent)
|
|
are inherently flawed.
|
|
|
|
One of many problems with this presentation
|
|
is that John is described as young, inexperienced,
|
|
and of low intelligence.
|
|
John is a typical lucky fool:
|
|
there is no indication that he has _any_ strategy,
|
|
much less that he is practicing [[modern-portfolio-theory]].
|
|
|
|
If I mistake Taleb's intent,
|
|
and the story was only meant to convey
|
|
that experienced, conscientious, and cautious decision-making
|
|
_can_ lead a person to success,
|
|
then I'm not sure who he's arguing against.
|