[Autism] Quantifying superficial differences between groups

Lets define the divergence between two groups of people, group A and group B, as the expected value of the number of people you would see from group A and group B before you could tell with 75% accuracy which group was which. 

So,

D(Goldman Fig, KKR PE) = 3    -> If I gave you 3 random ppl from GS and 3 random ppl from KKR PE, you could distinguish which group was which with 75% accuracy

Here are some values

D(all PE with age in 20s, all IB with age in 20s) = 6
D(MS Healthcare, MS TMT) = 25 ( I think there are demographic differences)

D(MS Fig, GS FIG) = inf?