Estimating the IQ of Gary Kasparov
Kasparov underwent intelligence testing conducted by the Spiegel magazine - and scored 123 on the ravens and 135 on another assembelled by Eysenck. He is also one of the best chess players who are currently alive - conservatively estimating he’s the 5th.
Currently assuming a g-loading of 0.7 for the ravens, a g-loading of chess of 0.4 and g-loading of 0.9 for the other test.
It’s too difficult to find somebody as extreme as Kasparov in chess performance with these particular IQ scores, so I’ll report the results of the simulations restricting the simulated Kaparov’s to a certain degree of chess performance (all have SEs of roughly 5.2).
Restricted to -1+ z: 132 (SE 6)
Restricted to 0+ z: 132 (SE 6)
Restricted to 1+ z: 132 (SE 6)
Restricted to 2+ z: 133 (SE 6)
Restricted to 3+ z: 134 (SE 5.8)
It seems that restricting the amount of chess skill doesn’t affect the estimates much, given that he already took two IQ tests. Based on the fact that he has a chess z-score of roughly 6, I would estimate he has an IQ of 135 with an SE of 5.7.
########################
set.seed(1)
g <- rnorm(60000000)
set.seed(2)
iq <- 0.9*g + rnorm(60000000)*sqrt(1-0.9^2)
set.seed(3)
c1 <- 0.7*g + rnorm(60000000)*sqrt(1-0.7^2)
set.seed(4)
c2 <- 0.4*g + rnorm(60000000)*sqrt(1-0.4^2)
subby1 <- data.frame(iq, c1)
subby1$g = g
subby1$c2 = c2
#change c2 to change chess skill minimum
subby2 <- subset(subby1, (subby1$c1 > 1.4 & subby1$c1 < 1.66) & (subby1$c2 > 3) & (subby1$iq > 2.2 & subby1$iq < 2.46))
mean(subby2$g)*15
sd(subby2$g)*15