In the comment section of Steve Sailer’s blog, the widely admired Ron Unz said the following about physicist Richard Feynman’s reported IQ of about 125:

…Like I said, Feynman was a notorious prankster and jokester. I’m somewhat more inclined to believe he was just pulling that journalist’s leg, and the silly fellow didn’t realize it.

But my main point is there’s simply *no* possibility Feynman had an honest-to-goodness true IQ of 125. 251 or maybe even 512 would be likelier. Back in JHS or HS he developed a new type of generalized Calculus after someone gave him a book. In college, he won the Putnam as a last-minute substitute entrant.

IQ-fetishists have come up with the silliest theories to justify the 125. Maybe he had a bad day. Maybe he did badly on the Verbal questions. The bottom line is if Feynman really scored 125 on a legitimate IQ test (a test that allowed scores higher than 125) then “IQ Is Bunk.”

I personally subscribe to the theory that Feynman’s verbal IQ was much lower than his spatial/mathematical IQ, and since prominent IQ tests of that era emphasized verbal ability almost exclusively, the test dramatically underestimated Feynman’s overall IQ.

However even if 125 were a true reflection of Feynman’s overall IQ, I’m not convinced the score invalidates the concept of IQ, because Feynman is only one data point. Unz seems to believe that if you’re a mathematical Genius, then by definition, you’re spectacularly brilliant. Maybe. But in my humble opinion, that’s analogous to arguing that if you’re a spectacular basketball player, then by definition, you’re super tall.

So just as height is perhaps the single best predictor of basketball achievement, intelligence is the single best predictor of mathematical achievement, but that doesn’t mean other variables can’t compensate for less than stratospheric height and intelligence respectively.

An example is Muggsy Bogues who became one of the best basketball players in Americas despite being only 5’3″ tall. 5’3″ is roughly 2.71 standard deviations below average in height for an American man; equivalent to an HQ (Height Quotient) of only 59!

By contrast, Feynman’s reported IQ score was 125 (1.67 standard deviations above the average American). So even if his true IQ were “only” 125, Feynman was more than FOUR STANDARD DEVIATIONS more gifted in intelligence than Bogues is in height. And yet no one cites Bogues to argue that tape measures must be bogus, even though he’s much much much more anomalous for basketball than Feynman was for physics.

Statistically we should expect some of the greatest intellectual achievers in history to have IQs below the gifted level. This is because achievement is an imperfect measure of talent, and talent in a specific domain, even one as g loaded as physics, is an imperfect measure of general or overall intelligence.

While the average Nobel prize winning academic probably has an IQ around 150, the scores are probably normally distributed, with some being above 180 and others being below 120. Analogously, we should expect the average height on the NBA to be about 6’7″, with some being above seven feet, and others being below 6’1″.

But again, I do not in any way believe that Feynman’s reported IQ score was valid, however I do believe some non-brilliant people achieve at the highest level in even the most intellectual of fields. That’s just the nature of an imperfect correlation between ability and achievement.