Estimating the IQ of Destiny

• 131 on mensa norway 6 months ago

• 130 on mensa norway 5 years ago

• LoL - tends to rank Diamond (2 SD above mean)

• Sc2 - peaked GM (tended to hover high masters on stream IIRC)

• 153 on the BRGHT.

There's not good data on the amount of Sc2 accounts at the time he hit GM - some people say it's about 250000, which seems like a fair estimate. This corresponds to a z-score of 4.3 - given that he mostly hovered around masters I would say it's a fair estimate.

He got pretty good scores on the online ravens tests, but those kind of suck, as ravens only correlates with IQ at about .7, and those particular tests were normed relative to a population of online test takers, who are probably above average in IQ. I have no idea how good the BRGHT is, but I doubt it’s great:

• 113 seems way too low for the average MIT student. Probably a non-negligible number of false reports here, but even with them, it does put the validity of the test into question.

• The rank order of IQ by major looks legit, and so do the countries (with some notable outliers)

• I took the test and the percentile was identical to the one I got on my 2nd SAT and on the old SAT.

Given that the test takes 20 minutes to take, I doubt its reliability is above .6. The opensychometrics IQ test and the WAIS correlate at .54 - it would be fair to assume the same reliability for the BRGHT.

Given he already has so much cognitive testing data, I’ll only take his Sc2 rank into consideration when estimating his IQ.

Inputting all of the numbers into the simulation suggested an IQ of 141 and an SE of 8.9.

set.seed(18)

g <- rnorm(60000000, mean=0)
iq <- 0.7*g + rnorm(60000000)*sqrt(1-0.7^2)
iq2 <- 0.7*g + rnorm(60000000)*sqrt(1-0.7^2)
lol <- 0.54*g + rnorm(60000000)*sqrt(1-0.54^2)
sc <- 0.45*g + rnorm(60000000)*sqrt(1-0.45^2)

subby1 <- data.frame(iq, iq2)
subby1$g = g
subby1$lol = lol
subby1$sc = sc
subby2 <- subset(subby1, (subby1$iq > 1.8 & subby1$iq < 2.3) & (subby1$iq2 > 1.8 & subby1$iq2 < 2.3) & (subby1$lol > 45/15 & subby1$lol < 60/15) & (subby1$sc > 3.5 & subby1$sc < 5))

#eliges <- subby2$g
eliges <- append(eliges, subby2$g)
mean(eliges)*15
sd(eliges)*15