Cultural Caviar

Negotiating the Curve

December 26, 2018

Negotiating the Curve

Nassim Nicholas Taleb is a prodigious intellectual who has managed to build a significant following with his books on investing and “tail risk,” such as The Black Swan and Skin in the Game. Yet he’s never quite gotten Watsoned out of polite society for his variegated crimethinking. All this despite a cantankerous personality, especially since he has taken up weightlifting and the testosterone has really gotten flowing.

Born in Lebanon to a ruling-class Christian family that was hurt by the horrific civil war of 1975–1990, Taleb, constantly worrying about what could possibly go disastrously wrong, is by nature a man of the right.

In contrast, Taleb’s intellectual archrival, rock-star Harvard psychologist Steven Pinker, an optimist who believes that every day in every way the world is getting better and better, would appear to be innately a man of the left.

But things are more complicated than that. Pinker, for example, is an outspoken advocate of the politically incorrect science of IQ and heredity. For instance, Pinker tweeted earlier this year:

The Blank Slate is cracking: With polygenic scores corroborating twin & adoption studies in showing IQ is in good part heritable, even schools & left-leaning mags are walking back the tabula rasa.

How does Pinker avoid getting in trouble like DNA researcher James D. Watson or Pinker’s friend Larry Summers, former president of Harvard until he gave a Pinkerian talk on sex differences in IQ? I’m not sure, exactly. Perhaps it’s that the lithe, long-haired, soft-spoken Pinker seems like the archetype of the liberal college professor.

Or perhaps would-be SJW deplatformers intuit that they’d come out of a collision with the extraordinarily smart Pinker as badly as Malcolm Gladwell did in 2009 when the New Yorker writer tried to taint Pinker with guilt by association with me and my impolitic views on IQ and race. Gladwell’s career has never fully recovered from the drubbing the seemingly mild-mannered Pinker gave him.

In contrast to Pinker, the brawny, bald, and obstreperous Taleb advocates largely PC views on the most career-endangering questions of race and IQ. It would certainly be prudent to not risk a career that has gained considerable momentum over the past decade, making Taleb one of the highest-paid highbrow authors in the world.

But perhaps Taleb has more intellectual reasons. He broke through to best-sellerdom in 2007 with The Black Swan. The title is Taleb’s rather strained and confusing metaphor for rare but important events. (A more straightforward title would have been the one Gladwell later used for his 2008 best-seller Outliers.)

Taleb argued that bell-curve-based forecasts often fail because outliers—“black swans”—can have more dramatic impact than expected under a normal probability distribution.

Taleb drew a helpful distinction between “Mediocristan,” situations where data points tend to be distributed according to bell curves (e.g., height, weight, IQ, or car accidents), and “Extremistan,” where rare items greatly affect the average (e.g., wealth or book sales per author).

After the financial crash, The Black Swan was acclaimed for its insight. But wasn’t the mortgage meltdown that triggered the Great Recession a classic instance of Mediocristan? It wasn’t set off by, say, Trump Tower defaulting, but by a huge number of moderate-size defaults in the exurbs of heavily Hispanic states, following a Bush Administration campaign to boost nonwhite homeownership by 5.5 million by easing down-payment requirements. The big losses weren’t in, say, Extremistan Malibu but in Mediocristan Inland Empire.

“It’s almost as if the IQ glass is somehow both half empty and half full at the same time…”

Attacking bell curves has been popular since the 1994 publication of Richard Herrnstein and Charles Murray’s IQ book. But, of course, they named their book The Bell Curve precisely because IQ is firmly lodged in Mediocristan.

Yet this didn’t stop Taleb over the weekend from launching a Twitter war on all things related to The Bell Curve: IQ, race, and genetics. Taleb began, curiously enough, by pooh-poohing the importance of tails:

1- What people don’t get about “race” ≠ or “heritability” of traits: even it these were true & not fabrications, they wd be dominated by IDIOSYNCRATIC differences. If pple from Mars had an “IQ” lower than earth by 5%, w/variance you wouldn’t see it in ANY 2-paired individuals.

Taleb added:

“IQ” is an arbitrary metric (say like length of the index finger), overoptimized, predictive of school grades, success as bureaucrat but not in the wild/entrepreneurship.

Taleb has made a lot more money than I have, so I’d listen to his views on entrepreneurs. On the other hand, a guy who worked for me in 1986 got a much better job the next year at Microsoft. He became a close acquaintance of Bill Gates when the entrepreneur wanted his girlfriend to go skydiving with him and she said she’d only do it if he found another woman to jump out of the plane with her. Gates looked over the roster for somebody ambitious and married and settled upon my old friend, much to the distress of his poor terrified wife.

My pal testified that Gates was intensely smart.

But the irony is those focused on the IQ of “races” (“bell curve”) seem to have a lower IQ scores themselves!

I wouldn’t say that Professor Taleb is right in the case of Pinker or Murray (who called Taleb’s tweet storm “willfully stupid”), but no, my IQ isn’t as high as his. I’m not as smart as Taleb, so the only way I can out-argue him is to use better facts and logic. (Here’s my 2007 frequently asked question list on IQ.)

In response to Taleb, I pointed out some obvious examples of racial gaps in real-world performance. For example, all 72 finalists in the men’s 100-meter dash in the past nine Olympics, going back through 1984, have been at least half black by ancestry.

Implies moderate race difference in sprinting speed piles up to huge race gaps in @nntaleb’s beloved tails.

Similarly, all starting cornerbacks in the NFL have been black since Jason Sehorn retired after the 2003 season. That’s two cornerbacks times 15 seasons times 32 teams, or 960 out of 960 data points.

That probably didn’t happen by luck.

Initially, Taleb tried rejecting any and all examples from sports.

Ludic. It has no impact when someone is crossing the street. Noise dominates outside of extremes. Try another example. Something nonludic.

But I stood my ground:

It’s silly to dismiss out of hand learning from sports statistics on grounds that they are “ludic.” E.g., the men’s 100m dash in the Olympics is an extremely global event with 100+ countries trying to get in….

Taleb eventually conceded that sports were a reasonable data source:

I accept that sports matter and draw a lot of people. And that heredity matters for some functions. But the tail is well…the tail. 90% and 99.9% of life lie by definition outside of the near and far tail respectively.

But of course sports success is Extremistan. Sprint king Usain Bolt makes $31 million per year and barely has to work out. When a journalist asked how fast Bolt could run a mile, his agent responded, “Usain has never run a mile.” In turn, what happens in Extremistan affects Mediocristan. I responded:

Sorry, @nntaleb, but race differences turn out to be highly relevant in real world because society is interested in extremes. E.g., all regular NFL cornerbacks after 2003 have been black: 960/960 data points. If your son is white, tell him to play safety, LB, receiver, not CB.

One interesting question is, how big is the racial gap in sprinting speed at the Mediocristan playground level? Unfortunately, I’m not aware of any studies of children’s running ability by race. The French had a project to time as many 15-year-olds as possible at running 50 meters, but they don’t record ethnic data. (Interestingly, the fastest white 100-meter man of all time, Christophe Lemaitre, was discovered by this national undertaking, which suggests that there may be a few other world-class white sprinters undiscerned by the usual methods.)

Taleb assumes the gap in sprinting speed between the average black and white child is trivial, but in my experience at playgrounds, it’s noticeably large. That has not insignificant real-world influences. I responded:

Sorry, but the extraordinary race gap in Olympic 100 meter dash results (all finalists since 1984 have been black) correlates with lots of mundane realities like: Who bullies whom in playground, and who takes up street crime: those most likely to outrun their pursuers.

It’s often lamented that black youths overinvest in the slim hope of becoming a professional athlete. But on the upside, it can make sense for a black child with above-average IQ and above-average foot speed to work hard on sports because the racial gap is large enough that a scholarship to a good private high school is hardly implausible.

For example, the one African-American in my Catholic high school class of 1976 was there on scholarship because as a senior he would run a 9.5 second 100-yard dash, the second-fastest in Southern California that year. He was recruited by UCLA’s track team, and today is an architect.

Nowadays, Catholic high schools recruit quite a few black athletes for each class.

And college scholarships are a possibility. A high school friend of my son’s got a free ride to Harvard because he is a giant black basketball player. He actually didn’t like sports all that much; he just worked extremely hard at his game after he grew tall so that he would be recruited by a top college, such as Harvard.

In contrast to sprinting speed, where we have ample data about the racial gap at the high end but not at the playground level, we know all about the IQ white-black gap in Mediocristan, but less so in Extremistan.

The notorious IQ gap is usually estimated at about one standard deviation or 15 points. The typical white would, by definition, score at the 50th percentile of the white distribution, while the typical black would score at about the 16th percentile of the white distribution.

But please note, this means that about one-sixth of blacks are smarter than the average white. Perhaps there are 7 million African-Americans with IQs above the white mean. So in IQ Mediocristan, the race gap, while sizable, is not immense.

On the other hand, in IQ Extremistan, there aren’t many blacks, although how few there are tends to be kept a secret. America is obsessed with pushing blacks to achieve high in many fields, but the numbers are in short supply.

For example, Penn law professor Amy Wax caused a huge brouhaha earlier this year by pointing out that blacks “rarely” finish in the top half of their class. The dean shot back that, in fact, “Black students have graduated in the top of the class at Penn Law.” But no data were forthcoming.

Amusingly, when Barack Obama showed up at Harvard Law School in 1988 and demonstrated that he could hang intellectually with the white kids, he was, within weeks, talked up as a possible first black president of the United States.

Taleb then shifted gears from race to IQ:

2- It takes a certain type of person to waste intelligent concentration on classroom/academic problems. These are lifeless bureaucrats who can muster sterile motivation. Some people can only focus on problems that are REAL, not fictional textbook ones.

I replied:

Actually, the military relies on heavily g-loaded cognitive tests such as the AFQT to determine who is allowed to enlist. By law, nobody below 10th percentile is allowed to enlist, and often Army/Marines take nobody below 30th %ile, & AF/Navy only recruit IQs of 100 or higher.

The military’s AFQT enlistment test isn’t precisely an IQ test, it’s more like the SAT/ACT college admissions tests, but it correlates closely enough with IQ that The Bell Curve used it as its main indicator of IQ.

I spent a couple of hours on the phone in 2004 with a college professor who had previously long been head of psychometrics for one of the main branches of the military and who had supplied the Pentagon’s test data that made up much of The Bell Curve. His only objection to the book was that Herrnstein and Murray hadn’t gone far enough.

During the Vietnam War, Defense Secretary Robert McNamara tried to get around the law banning enlisting soldiers from the bottom 10 percent of the intelligence bell curve. The book on what ensued is entitled McNamara’s Folly.

Taleb went on:

If many millionaires have IQs around100, & 58 y.o. back office clercs at Goldman Sachs or elsewhere an IQ of 155 (true example), clearly the measurement is less informative than claimed.

I said:

This is kind of like saying: if there are 6-footers starting in the NBA and 7-footers on the bench, clearly height is less informative than claimed. That’s true … to some extent, but it doesn’t disprove that height is a sizable advantage in the NBA.

Remember, there are lot more 6 footers than 7 footers, just like 100 IQs vs. 155 IQs. When I attended Rice and UCLA, there were a total of four 7 footers on campuses, all on the school basketball teams, and 2 played in the NBA, one (Mark Eaton) in NBA all star game.

It’s almost as if the IQ glass is somehow both half empty and half full at the same time…

This doesn’t mean that IQ is a perfect measure above criticism, just that in an imperfect world, it’s proved itself over the past 113 years as one of the social sciences’ enduring accomplishments.

Because Taleb has a much higher IQ than I have, the only way I can win an argument with him is by being right.

Pay to Play - Put your money where your mouth is and subscribe for an ad-free experience and to join the world famous Takimag comment board.