I'm a big Taleb fan, and have been for years, but he's not very effective in this format. That's as much a critique on the segment's director as anything. It's set up as a roundtable talking heads discussion/debate. (I assume most of the others' bits have been excised). His points are absolutely valid, but they require some ground work in how risk modeling has worked to understand them. Without that, he just seems to be sputtering somewhat inarticulately.

Maybe somebody with a media budget (you watching, PJ guys?) should sit him down with a competent interviewer and walk through the whole thing. You might also need someone as a foil to draw out and explain the modeling aspects to non-mathematicians. Taleb strikes me as an intuitive, and he's played with this material so long that his sense of the math has been baked him - he doesn't articulate it so well.

Tim: I suggest you drop Wretchard an email.

In terms of sentiment: He's a Cassandra, or a godless Jeremiah. He can't be right, people have gotten the Nobel Prize who disagree with him.

There's no huge profit to be made in acknowledging and acting upon the stodgy conservative consequences of his paradigm as an alternative to the shiny tree-ornament bauble of undeterminable fragility: the well-behaved Pascalian-Cartesian, distribution & variance, ludic fallacy. As the wind-up and punchline of the old joke go:

"Sheesh, lady, do you always have twins?!"

"Heck no, most of the time nothing happens."

These excerpts illustrate the huge gap. "Inside baseball" aficionados such as AL, Tim and myself can appreciate the wince AL points out. Many regular folks eager to be palliated will see only a crank.

"What is to be done?"

"Throw it all out. All the huge profit machinery contains incalculable risks."

"Oh, now, that can't be right."

Man is not a rational animal, he is a rationalizing animal.

The deep irony is that financial disclosures always say "past results are no guarantee of future earnings." But surely that's just a formality, right? Bad things can't happen to good people, can they? Surely all the system needs is a tune-up?

Magical thinking. Sentiment. Epistemic arrogance.

[slight edit]

PS: Thinking of Pascal and the origin of gambling theory / ludics led to this observation: The belief that we can keep throwing effort at Mediocristan while living in Extremistan is in effect a meta-Gambler's Ruin: Baby, my luck's gonna change, I can feel it!

Tim: I suggest you drop Wretchard an email.

Done, assuming his old hotmail address is still valid. If not, could you or AL send me an updated?

AL might have it, I don't. But I can use my Google-fu and see what's out thar' yonduh. Will advise.

All I have is the Hotmail one...

A.L.

PJM Editorial has an 877 number. Maybe AL in his copious spare time could wheedle a fresher contact from them, or at least give them a heads-up and then Tim or Marc could mail a message to be forwarded. I dunno. Seems like an obvious project.

Hey. Bill Whittle, anyone?

The core of the problem is that most economists believe that the extremes of risk operate according to Gaussian (bell curve) distribution, when the evidence is increasingly clear it is in fact a power function of some sort.

That's how you get several "once in ten thousand year" events in the space of merely ten years, not ten thousand.

Yes... Well, Gaussian or Gaussianoid (bimodal, or what have you). The key is that they believe in the tractability of the system using the tools they have, even though they can't run controlled experiments and there is no underlying deep model to inspect. No CERN accelerator to probe for the fiscal Higgs Boson.

And then they curve fit without taking into account how easy it is to fool yourself with some smooth path through a point cloud when you can't conduct experiments to test your models; and they rebake their models after every fiscal mishap or surprise; and each time they say look, we can backcast it pretty well, finally. No more big surprises. We can do this thing. We know what we're doing, we work really hard on this stuff.

Though their newest-improvedest backcasting still uses the same suspicious curvefitting with expectations of well-behaved distributions modeled by formulae and factors cooked up ad hoc.

Reminds me of Bullwinkle J. Moose and the rabbit in the hat. "This time, for sure!"

Though their newest-improvedest backcasting still uses the same suspicious curvefitting with expectations of well-behaved distributions modeled by formulae and factors cooked up ad hoc.

Recalling one aphorism from my long-ago statistics and systems education: With enough parameters you can fit an elephant

I want to add one, in the spirit of TANSTAAFL:

Backcasting Ain't Forecasting No Matter How Funny You Hold Your Face.

But that's not been primped into pronounceable form yet.

Hmm. For pronunciation of the above: "Baffnum-high-five?"

Alllmost.

I also wonder to what extent computers and calculators got us into this mess. I'm old enough to have done all my undergrad work, and much of my post-grad studies on a slide rule.

The true beauty of a slide rule is that it keeps you honest about significant figures.

In other words, we now have at least a generation of market quants who fundamentally do not understand the profound difference between precision and accuracy.

The digital thermometer reads a very precise 74.361 degrees F, but when you turn it over the small print says ±3 degrees!

So they cobble together a three-way cross-currency deal hedging the Argentine Peso--Japanese Yen exchange rate against the relative performance of India's Sensex versus the Nikkei, all allegedly underpinned by 180-day sovereign Argentine debt and a bunch of second-order calls against the MerVal close on 19 September, all of it geared out at about 20-to-1.

It's calculated to pay off big time on a positive swing of 0.02% ... except nobody noticed that Argentine economic data are not even accurate to ±2%. And we're surprised it all blows up?

Much of what has been happening is not so much the result of Taleb's beloved black swans as it is the normal and highly foreseeable product of the intersection of ignorance with hubris.

Taleb agrees with your last para -- he indicates that in the video.

Even so, the pundits and wonks will ASAP busily cook up a just-so story for how it all went down, ignoring details and still claiming more knowledge than they actually possess -- and that sort of rhymes with the pattern Taleb describes après-Swan too.

Ya know? I'm just a small business owner trying to protect my business, my customers, and my family from other people's pathological stupidity ... and cu -pidity. I hafta say I'm getting rather bitter about it all.

Another part of the underlying financial problem is that competition for higher yields led to the construction of some absolutely bizarre financial instruments for the simple reason that they couldn't be copied or commoditised by competitors.

Unfortunately, neither could those monstrosities be sufficiently analysed and rated by those charged with that task ... or even figured out by those investing in such Kafka-esque constructions.

Taleb is being polite.

Hmmm... Stu-cu-pidity. I like it. Probably won't catch on as a phrase.

#16 from Nortius Maximus at 6:12 pm on Oct 14, 2008
Hmmm... Stu-cu-pidity. I like it. Probably won't catch on as a phrase.

I have always been fond of the Triple B theory, explained to me very succinctly by a risk manager over dinner "Bullshit Baffles Brains."

THREE RULES FOR RISK MANAGEMENT
1. Don't put all your eggs in one basket
2. Never go all in
3. Diversify

I don't care what is on the table with the bailout, these elitist illuminati aren't thinking of my interest. I don't have the money to pay for someone else to get rich. If socialist share the wealth, why are they making the tax payers pay for the wealthy to stay rich?

I don't care what is on the table with the bailout, these elitist illuminati aren't thinking of my interest. I don't have the money to pay for someone else to get rich. If socialist share the wealth, why are they making the tax payers pay for the wealthy to stay rich?