« Underdogs | Main | More on Quarterbacks »



To be fair, our host called Sailer a racist openly, so what do you expect Sailer to do?

S Perle

"his belief that black people are intellectually inferior to white people."

It takes a certain level of intellectual dishonesty to misrepresent someone's views in this way.

As noted above, Sailer has simply reported on empirical facts - groups differ on average.

If two groups differed by, say, 10 points in average IQ (2/3 of a standard deviation), the respective distributions would overlap quite a bit (more in-group than between-group variation), but the fraction of people with IQ above some threshold (e.g., >140) would be radically different.

That is the case with Ashkenazi Jews, psychologists and educational researchers have pegged their average IQ at 107.5 to 115.

That’s only modestly higher than the overall European average of 100, but the gap is large enough to produce a huge difference in the proportion of geniuses. When a group’s average IQ is 100, the percentage of people above 140 is 0.4%; when the average is 110, the rate is 2.3%.


S Perle

If Pinker had wanted to really take Gladwell to task he could have pointed to the claim about asian math success and rice cultivation.

This culture explanation for math success is easily disproven by transracial adoption studies. If Asian academic success was really due to some special set of academic values inculcated by Asian parents (something not demonstrated by the data to begin with), then why do Asians do better academically than whites even when they are raised by white parents?

Transracial, same-race adoptions, and the need for multiple measures of adolescent adjustment. (Burrow, Anthony L.; Finley, Gordon E.
American Journal of Orthopsychiatry. Vol 74(4), Oct 2004, 577-583.



No one to whom popularity is important is willing to wade into the treacherous sea of IQ and race--unless, of course, like Gladwell, they toe the line dictated by polite academic society. To show any kind of interest in the subject is to risk one's academic and professional standing.



I think what you did not take into consideration is that late-round quarterbacks usually aren't good enough to get on the field in the first place. A coach selects his quarterback based who gives the team the best chance to win.

Obviously, the gap between a late-round quarterback who's good enough to get on the field and a early-round quarterback who's good enough to make the field isn't going to be that large because both quarterbacks have already met a certain threshold of ability. The more important question is, "How many late-round quarterback are capable of being an NFL starter?" The decisions of the coaches, who have the most to gain or lose, would indicate: "Not many."

Alan Tower

Dear Malcolm:
Dude come on. You do incredible work, you are someone who many of us just wait for the next fascinating analysis. You dont need to get defensive when someone points out how some of your research, like in the NFL malarkey might be off point, weighted and rests too much on some academics narrow view (like this is actually important in the world?) Why not just take it in and extend?

To imply the Sailer guy is in some way a racist( probably not accurate based on his work) and not dealing with his thoughtful analysis is letting your emotional center take over isn't it? What we learned from Outliers . . . You dont really own all this great success, you and everyeone else are all continguent beings on so many factors is makes the head spin. Right? So why bother with this kind of reaction in your blog?

What is coming through you in your approach to understanding and contributing to the world is quite astonishing. Thanks for being here for us.

If Mr. Pinker needs to do his thing to you in the way he sometimes does it elsewhere, why not let him and shine back on him who you really are, the hugeness of what you are. It might rub off on him and on all of us. We all then expand in the process.

Munroe Ross

I encourage you to take this criticism in good humour, because Pinker gets one big thing right: you need to know more about numbers, and it will improve your insights tremendously. And really you can't blame him for highlighting your weakness in this area. Given who he is, he couldn't not call you on it.

It is disappointing to me that you got signal detection theory wrong (which Pinker dinged you on), beacuse it's a thing of beauty that explicates so many things. Trust me, once you get it, you'll be in love with it.

Oh, and I'm also voting for the beer summit.

black sea

Of the 32 current starting NFL quarterbacks, the draft rounds in which they were selected are as follows:

Round 1: 18

Round 2: 3

Round 3: 2

Round 4: 2

Round 5: 0

Round 6: 3

Round 7: 1

Undrafted: 3

More than half of the starting quarterbacks are former first rounders. Regarding the 3 who were undrafted free agents, bear in mind that slews of players are signed as undrafted, much more than the 32 who are taken in any particular round.


If you take the 168 qbs drafted in the first 250 players from 1980 to 1999 (following Sailer's link to pf ref .com) and calculate the average length of their careers by draft position, you get this:

1-50 9.2 years
51-100 8.1 years
101-150 6.4 years
151-200 4.5 years
201-250 3.5 years

It's not far from what you'd expect, I think. This isn't the best stat but it was quick and easy to calculate. And I think it's an illuminating one: NFL teams generally don't give you a roster spot unless you're useful. Whatever might happen in a player's early years because of various biases due to draft position, contract, chances to play, quality of team, etc etc etc, I think it's hard to argue that "the position a quarterback is taken in the college draft is not a reliable indicator of his performance as a professional." Unless you're going to define "performance" in a way fundamentally different from the way NFL teams define it.

I don't know whether Berri and Simmons have found something or not; maybe the "qb yield curve" is flatter in the top 90 than one might expect. Still, the question is why? Why does an effect that clearly exists when comparing (say) draft slots 151-200 with 51-100 will vanish completely when you compare 11-90 with 1-10? The answer is it probably doesn't, but maybe there is something interesting going on with those 1-10 guys. Maybe it's as simple as bad teams taking undeserving qbs early. And remember the sample size is not enormous....

Clark Bow

As others have noted here, "igon value" is not a "spelling error." It's an admission that an author is willing to expound on a topic he has no knowledge of.


Rhetorically, good argument, but you make use of both ad-hominem and argumentum ad verecundiam, both logical fallacies. It may be, however unlikely, that a racist can be right about football, and that a fantasy football blog may be right about football too where a university study is wrong. You'd have done better to attack the fantasy football site's "findings" purely on their merits, and possibly go after Sailer for his racism in a separate post - although you're probably right you've not given a logical proof, and I'm confident you can easily enough.


While your rebuttal of Pinker's claims dealing with the (to use Pinker's term) cherry-picked matter of NFL quarterbacks is spot on, you fail to address any of his more legitimate qualms with your application of statistics in general.

We all know that catchy quips sell books, but I think your readers might be better off with less ketchup stories and more expected values, failure rate distributions, and ohohoho, eigenvalues. Luckily for you, our great country has something called "introductory statistics courses". Try it out sometime.


"100% classy and predictable, mr. sailer - comment and blog away into the void...4 comments in an hour? wow."

Posting extensive, lucid and fact-stuffed reasoning? How dare he!

Instead, let´s celebrate your more helpful contribution to the debate, which consists of...

...getting all snarky at those posting extensive, lucid and fact-stuffed analysis.



Well done Malcolm. You are not "minor genius" as prof. Pinker Sarcastically claims. You r simply genius.

paul schejtman

maybe gladwell doesnt think he is smart deep down.

Joe Torben

It's hard to be more spectacularly wrong than Mr Gladwell is on the success of QBs. You need to know absolutely nothing about any of the research except to know that the "per play" statistic gives it away. As we all know, NFL games are not won on a "per play" basis. They are won on total score accumulated.

Why, then, would anyone invent such a statistic? It should be obvious to anyone that the reason is that this particular skewing gives the desired result. A little more curiosity from Mr Gladwell would have revealed that instantly. And if not, he could have talked to any statistician. Or tried to understand opposing points of view.

The ad hominems in his rebuttal were embarrasing, and also a very strong clue about the strength of the argument.


Malcolm you've just been bombed by a blog-full of racist Sailerites with nothing better to do with their free time.

You kick-ass in my books. :)


From what I've read of your writing, I think Pinker's primary point (that you sometimes cite math that you don't fully understand) is somewhat valid. It's a bit disingenuous of you to call the "Igon Value" thing merely a spelling issue; anyone with even the faintest memory of introductory linear algebra would have taken a better stab at the spelling than that, particularly since the meaning of the word is somewhat implicit in its etymology. If I were to read any piece talking about "igon values," my reaction would be, "this guy had a half-hour minute chat with a mathematician and now he thinks he's competent to deliver a math lecture."

Other than that, I thought Pinker's article was cruel and petty at best, totally wrong and hypocritical at worst.

Moreover, from what I've read of your stuff and Pinker's (pop sci) stuff, your stuff wins. By a lot.


Malcolm, Igon Value is not a simple misspelling. If you were talking about Boston and wrote Masachusets, that would be a misspelling.

Writing Igon Value instead of eigenvalue shows you never bothered to try and find out what eigenvalue means. You would have otherwise realized that the spelling was completely off.

What really bothers me is that if you had written Sheck Speah when interviewing a British academic about Hamlet, you would have lost all credibility. Because most people are not exposed to this kind of math, writing "Igon Value" instead of eigenvalue may be passed as a simple spelling error.


Maple Leaf Beef !

Charlie A. Roy

Love your work and great rebuttal. I think you should take some of that same scholarly attitude and apply it to your take on KIPP and the whole time on task argument. I like your thinking but there is another side to KIPP that is often ignored in the media. Of course I'm taking points primarily from Jim Horn's "School Matters" blog so maybe I should do a little more of my own scholarly research. Be well and thanks for always providing such thought provoking work. I love your latest release.


As a scientist, I don't let the occasional odd sentence bother me. Indeed, it's the sort of thing that makes writing resonate, and I'm sorry that David Foster Wallace is not around to appreciate the idea of an Igon Value.


The issue is not one of spelling but of understanding. If you are attempting to argue the merits of some mathematical models of the stock market versus others, you had better have some cursory understanding of the mathematics behind those models. "Igon value" is just one particularly revealing instance of your lack of that understanding. Instead of attempting to explain the mathematics in terms people can understand, you quote these people as if you are in complete awe of their understanding that you lack, and the mathematics comes across as this esoteric gibberish that nobody should ever be expected to understand.

You also vastly oversimplify the small amount of mathematics that you do attempt to explain. Nobody actually thinks the stock market follows a perfect Gaussian distribution. However, anything that involves a bunch of independent, nearly identical events or decisions can be approximated by a normal (aka Gaussian) distribution. The stock market involves many decisions made by many people in similar situations, and in a short span of time, these similar decisions can be considered as independent of one another and identically distributed. For example, if there are one million people in similar financial situations, each of whom make the decision to hold or sell one share of stock in some particular company, each person will have essentially the same probability of holding or selling as any other person, and thus the overall behavior of those decisions can be described by a Gaussian distribution. Then on top of that, interactions between the actions of different people can be taken into account, and then the effect of potential external stimuli on the market. No one claims that over a long period of time the stock market behaves normally.

I am not an expert in modeling the stock market, and this is just my quite possibly flawed understanding of how a little bit of it works. However, you seem to make no attempt at explaining, or even understanding, the mathematics behind the different theories, instead treating it as the privileged information of the few who understand it. I assure you that with your access to so many of these quants, at least a few of them can explain some of what they do in terms a lay-person can understand. Instead, you just overhear them talking in terms you don't understand, hence "Igon value".

(For your information, here is a way of understanding what an eigenvalue is. A linear operator is something that takes one vector in n dimensional space (i.e. an ordered list of n numbers) and applies a simple rule to give another vector in n dimensional space (in particular each coordinate of the new vector is a specified linear combination of the coordinates of the old vector). For some vectors, applying this linear operator is the same as just going through and multiplying each element in the list by a specified number, lambda. Then that vector is called an eigenvector of the operator, and lambda is called an eigenvalue.)

Ray Sawhill

If Malcolm wants to show how truly resourceful and courageous a journalist he is, he'd write a piece about Steve Sailer -- go meet him, read his writing carefully, wrestle with its substance and the research it builds on, and look into (and give some thought to) the way Sailer has built up a following and has become (in a not-publicly-acknowledged way) quite influential.

I don't know about anyone else, but in recent years I've had numerous moments, when reading mainstream coverage of various hot and dicey topics, when I've found myself thinking, "Y'know, I bet this writer has been following Steve Sailer -- and wouldn't dare admit it in public."

The Sailer phenomenon is a good news story, no doubt about it. Go for it before someone else does.


It's not that IQ above 120 doesn't matter, it's that IQ above 120 doesn't predict achievement very well. Sure, all the major scientists are smart. What about all the smart people who didn't become major scientists, or otherwise high achievers?

The thing is (I have been taught), the scatterplot of IQ vs. achievement is heteroscedastic. At lower IQs, the relationship is strong. At higher IQs, it's not.

The comments to this entry are closed.

My Photo


  • I'm a writer for the New Yorker magazine, and the author of four books, "The Tipping Point: How Little Things Make a Big Difference", "Blink: The Power of Thinking Without Thinking" and "Outliers: The Story of Success." My latest book, "What the Dog Saw" is a compilation of stories published in The New Yorker. I was born in England, and raised in southwestern Ontario in Canada. Now I live in New York City.

    My great claim to fame is that I'm from the town where they invented the BlackBerry. My family also believes (with some justification) that we are distantly related to Colin Powell. I invite you to look closely at the photograph above and draw your own conclusions.

My Website


  • What the Dog Saw

    buy from amazon


    buy from amazon

    buy from amazon UK


    buy from amazon

    buy from amazon UK

    Tipping Point

    buy from amazon

Recent Articles

Blog powered by Typepad