I have the greatest book EVER

[Cain]

I know it helps validate some of the key arguments in my dissertation.  And make me a complete asshole (according to everyone else) in International Relations. 

Not that this is exceptionally hard.  For example, the main allegedly predictive theory in IR is Neorealism.  Yet it failed to predict the end of the Cold War, which some would say is a pretty big failure of the theory.  Neoliberalism is another one, which has failed to explain the anti-globalization movement and the shift in South America towards left wing populism.  In short, there are all sort of wannabe scientists running around in policy circles, taking their talking points from theories which only bear somewhat of a resemblance to reality.

[Epistemologicus Boobicon]

This sorta vaguely ties into something I was reading while on the bus to work this morning--Murray Rothbard's Making Economic Sense. Not very Discordian, I know, but he had this essay in there that he wrote back in the early 1980s. In it he talked about how the entire statistics profession was undergoing a radical transformation at the time because there was finally a refutation of the "normal curve"---refuted by "artificial data sets" generated by computers. Now that struck me as Discordian.

[Triple Zero]

reading further into this book, i already commented on Taleb's tendency for "funny" remarks, but it seems that this guy isn't really going to stop making it very very hard to take him serious, is he?

Let's play the following thought experiment. Assume that you round up a thousand people randomly selected from the general population and have them stand next to one another in a stadium. You can even include Frenchmen (but please, not too many out of consideration for the others in the group), ...

c'mon is he for real? talking about such interesting subject matter, and at the same time pissing me off with such dumb "funny" remarks.

makes it harder for me not to instantly disagree with whatever else he has to say, bringing me to the following:

Matters that seem to belong to Mediocristan: height, weight, [ list of random variables and phenomena with--or based on--a Normal distribution].
Matters that seem to belong to Extremistan: wealth, income, [ list of random variables and phenomena with an exponential/power law distribution]. The Extremistan list is much longer than the prior one.

what, one is longer than the other? you can think up an infinite list of examples for both classes of random events, so what exactly does he mean when saying that one is longer than the other?

in his list of differences between class 1 (Normal) and class 2 (exponential/power law) events:

[Class 1 / Mediocristan is] more likely to be found in our ancestral environment
[whereas class 2 / Extremistan is] more likely to be found in our modern environment

didn't he just explain that power law distributions have already been going on since evolution started storing information using DNA? and didn't he just say that America specializing in the creative aspect of things, exporting their culture according to class 2, heavily depends on outsourcing the (class 1) grunt work to other countries (and not only developing ones)? so how is either class more likely to be found in either era.

i find this very funny, because the guy seems to have fallen for his own traps:

This framework, showing that Exremistan is where most of the Black Swan action is, is only a rough approximation--please do not Platonify [= categorize / polarize] it; don't simplify it beyond what's necessary.

and i'm looking forward to his justification of using the terms "fractal" and "mandelbrotian" (lolwhut), if any

another quote, still the same chapter and it's only getting worse:

Extremistan does not always imply Black Swans. Some events can be rare and consequential, but somewhat predictable, particularly to those who are prepared for them and have the tools to understand them (instead of listening to statisticians, economists, and charlatans of the bell-curve variety[1]). They are near-Black Swans. They are somewhat tractable scientifically [...] his category encompasses the randomness that produces phenomena commonly known by terms such as scalable, scale-invariant, power laws, Pareto-Zipf laws, Yule's law, Paretian-stable processes, Levy-stable and fractal laws[2]

[1] did he just really say that economists and statisticians do not know about power-laws??!! come on please, the buzz about "leveraging the long tail", "the 95/5 rule", etc * pre-dates this book by at least several years, and scientific research into these distributions by several decades. why else would these events be "commonly known" by big names such as Zipf and Levy? perhaps not as well-known as the Gauss distribution, but any statician not knowing these distributions would indeed be a charlatan, IMO.
(OTOH, who knows, maybe it *is* that bad, i just happen to know most of these terms from my general interest in the subject, i will ask a friend who majored in statistics, i bet he would love this book too, if mr Taleb could stop calling statisticians charlatans)

[2] i'm a littlebit familiar with about half of these terms, and afaik, they all mean kind of the same thing in this context. and just about every quantifiable example (that is, relating to numbers) Taleb has given for Black Swans falls into one of these categories (that he just called near-Black Swans!). with the specific example of "uses of words in a vocabulary" which is basically what Zipf's law is about. in fact, when he defined Black Swans a few chapters earlier, i immediately had to think of power laws. so if those are near-Black, or "gray" swans, then what is a real Black Swan? even more random, unpredictable? well, colour me intrigued.

it seems to me that mr Taleb has no problem at all making fun of statisticians, but doesn't really care get his facts straight or be a littlebit exact about the terms he's throwing around.
economists i don't really know about, but the ones that i know who concern themselves with matters of risk-analysis have majored in either math or physics/astronomy and know their shit--but maybe in America they're too busy being "far, far more creative than these nations of museumgoers and equation solvers". this may sound like i'm just being annoyed because he dissed Europeans (and i must admit, it did tick me off a littlebit), but this guy just defended being ignorant about math and equations, while on the other hand complaining about statisticians and economists being (apparently) ignorant about math and equations... so apparently a statistician or economist is either "uncreative" "equation-solving" "snotty (and frustrated)", or, indeed, a charlatan. can't really disagree with that, now can i?

i'm still of the opinion that mr Taleb has indeed something new to tell, with the idea that (near?) Black Swans have such an enormous impact on the history of things that Normal/Gaussian/predictable events pale in comparison. as far as i know, the consequences of power law and exponential distributions have not yet been approached from a historical angle, and indeed his findings in that respect are quite interesting. i just wish that he would keep his arrogance to himself, especially when it causes him to be wrong or contradictory.


* not surprisingly, this buzz coincides with the buzz about "web2.0", which, contrary to the gut-reaction of most IT people, is not complete bullshit, but just a very ill-defined term, meaning a lot of different things ranging from marketing bullshit like shiny logos and websites named like Star Wars characters (meebo.com?) but also some interesting observations and developments on the subjects of social networks, hierarchical (directory) versus horizontal (tagging) clustering, online communities and the probability density distributions therein.

[LMNO]

I'll agree he's a bit of an ass, but I think he's using it as a device, a way to keep the text "fresh" so it doesn't look like a textbook.  I don't think he's serious about his euro-bashing, considering he's from Lebabnon.

I'm still only about halfway through, so right now it only looks like he's tearing everything down.  I'm waiting for him to build it back up.  If he doesn't, I dunno.  Another case of pointing to problems without any solutions, I suppose.


Although, it's interesting that in the light of "Black Swan", all of a sudden Donald Rumsfeld's "known knowns/known unknowns/unknown unknowns" speech actually becomes quite astute.

[Cain]

I found it added some levity to the book.

And further confirms my suspicions that he is really a Discordian.

[Cramulus]

good critique, zilch. It does sound like you're mostly annoyed by his personality and his disgust with... well just about everybody, right? He doesn't like statisticians, he doesn't like economists, and IIRC he sort of tolerates other philosophers. I got a similar sort of vibe from Talib. I think I called him an arm-chair philosopher earlier in this thread, though Cain corrected me- he does have quite a bit of in-field experience in risk-analysis. I felt like the book was pushing at "No one really knows WTF is going to happen, and the more certain they are, the more shocked they're going to be. Idiots."

He doesn't try to make friends. He favors Americans over Europeans (because Americans are supposedly "more creative" than Europeans), and people from the Bronx over people from Manhattan (for their street-smarts). For a New York Intellectual, he spends quite a bit of time separating himself from them!

and jah I confess I only got about halfway through Black Swan before I was distracted by other IRL issues. When I lost track of it, I felt like Talib had been over-rehashing for about a hundred pages.  I found it fascinating, but it wasn't moving quick enough to keep my attention. Should I pick it back up and push through the end?

[Cain]

I like his iconoclastic attitudes towards intellectuals - mostly because I share them. 

He gets a little more mathematical and economical after chapter 15.

[Triple Zero]

(LMNO just FYI, Lebanon is not in Europe, plus he moved to Pennsylvania not after '82)

if it was just the personality i could live with that. also, the jokes wouldn't have been annoying at all if they weren't wrong to begin with

in my post above, i pointed out some things where he was simply wrong, inaccurate or contradictory. now jokes and attempts to keep a text "fresh" i can deal with, but making statements that are false just piss me off when reading a book (especially when he's talking about a subject that, in principle, i'm inclined to agree with).
and really, i don't mind the euro-bashing, i frowned at it once, and silently replaced the word "Frenchmen" with "Mexicans", and assumed he was talking about South-America versus Asia :-P either way, my point about "equation-solvers" versus "charlatans" still stands.

the things he assumes statisticians do and do not know about, i find questionable.

now, i don't want to debunk this book, as i said, i love the subject, but please consider the next quote:

Statisticians, it has been shown, tend to leave their brains in the classroom and engage in the most trivial inferential errors one they are let out on the streets. In 1971, the psychologists Danny Kahneman and Amos Tversk plied professors in statistics with statistical questions not phrased as statistical questions. One was similar to the following (changing the example for clarity): Assume that you live in a town with to hospitals--one large, the other small. On a given day 60 percent of those born in one of the two hospitals are boys. Which hospital is it likely to be? Many statisticians made the equivalent of the mistake (during a casual conversation) of choosing the larger hospital, when in fact the ver basis of statistics is that large samples are more stable and should fluctuate less from the long-term average--here, 50 percent for each of the sexes--than smaller samples.

personally i found this just a littlebit hard to believe, given that IMO, at least anybody with a college-level highschool diploma and math/statistics in their final year would answer this question correctly. also i really wondered in what way you'd have to phrase the question in order to confuse statisticians about it.

so, i looked up the original article. and apparently:

- they weren't "professors in statistics", but they were people attending two symposia for psychologists.
- they weren't "questions not phrased as statistical questions", they were questions about scientific experiments, regarding confidence intervals, sample variation, Student's t-test and other technical statistical things.
- "changing the example for clarity" seems to mean, completely making something up. there is nothing even remotely like the "quoted" example in the original article.

now, at least over here, psychologists aren't exactly known for their mathematical knowledge and insight. of course they should, because statistics are very important to their area of research. but in general they aren't(even though i know a few examples to the contrary). indeed, if Taleb would have said that psychologists are charlatans because they (generally) don't know shit about statistics, i wouldn't have disagreed as vehemently with his statement as i did when he spoke about statisticians.

btw, i don't want to insult any psychologists that happen to read this board. a friend of mine, graduated psychology a few years ago and kicks ass at statistics. on the other hand, my father is a respected neuropsychologist, he's pretty good at his job, visits the sort of symposia this questionnaire was taken at, and would definitely, totally flunk it.

the article itself isn't completely free of errors, wrong assumptions and false implications either. apart from my own doubts i found at least one review that debunked the conclusions drawn in the article.

(and, i'm not really sure what to make of the fact that Taleb refers to Daniel Kahneman as "Danny")

now, i just found this out because i happen to like checking sources once in a while. it took me about an hour to find and read the article. can you now understand why this pisses me off severely? i don't have the time to check for every reference, every example Taleb gives whether it is true, or if it's completely made up bullshit irrelevant yet vaguely related to the topic at hand, as in this case.
sure, i'm gonna finish the book, i enjoy it too much, but i'm a littlebit scared, at least with one story--i won't tell which one, avoiding spoilers--he admitted it was a joke (a good one at that, i enjoyed it, although it also didn't help my nervous skepticism).

he does sound discordian enough to me (also the story about the pranks), and what are discordians to do to eachother, if not disagree with one another?

[Cain]

Actually, historically, the Levant was considered part of Europe.  You see, you had the Byzantium Empire and Crusader Kingdoms all the way down to the modern Egyptian border, and even after they were eventually pushed out by the Ottomans and Saladin, they were still considered Christian, and thus, European, territory.  Its only been with the ascendancy of the Shi'ite 'minority' and the Civil War the Levant has been considered Middle Eastern.

[Triple Zero]

either way, deliberately misquoting scientific articles in order to support an argument is a capital sin in my book.

[LMNO]

000, what I think he was getting at was that humans who are "experts" in something often turn the "expert" way of thinking on and off.

So, a professional statistician won't always think like a statistician when not it statistician mode.

As a musician, I can appreciate a good pop song, and can find interesting melodies and lyrical turns of phrase in the most trite of Top 40.

But that doesn't mean when I hear the new Brittany Spears single I don't say, "turn that shit off!"

[Triple Zero]

yeah, except that is not entirely true. i know a professional statistician, and he definitely has a clearer grasp on seemingly obscure risks and probabilities even outside of the classroom.

also, the article he "quoted" to make his point was about psychologists, not statisticians. as i said, if he would have ranted against the statistical knowledge of the average psychologist, i wouldn't have had a problem with it at all.

[Cramulus]

BTW Big Kudos, tripzip, for checking his references. That's some sterling academic professionalism right there.

It feels "truthy" though, no? 

[LMNO]

"One was similar to the following:"


He's making a parallel comparison.




And isn't your anecdotal evidence of "a statistician you know" just as suspect?

[Bebek Sincap Ratatosk]

000, what I think he was getting at was that humans who are "experts" in something often turn the "expert" way of thinking on and off.

So, a professional statistician won't always think like a statistician when not it statistician mode.

As a musician, I can appreciate a good pop song, and can find interesting melodies and lyrical turns of phrase in the most trite of Top 40.

But that doesn't mean when I hear the new Brittany Spears single I don't say, "turn that shit off!"

But there's a big difference between "what he's getting at" and the example he's using to prove his point. What he seems to be aiming at is "a limitation of the human to apply knowledge of models in the real world"... but the example appears to be more about how most people don't really grasp the models related to statistics correctly. The former would seem important in the concept of Black Swans (the models can't really reflect reality), the latter would seem to counter the argument, in that with education about the models, the participants could perhaps have answered the question correctly. That seems like a pretty big misdirection to me....