Now let's do some math

With regards to the suspected condition of Caster Semenya and other Olympic athletes whose...crotches appear female, but their athletic performance appears to be male...one of the likely causes is something called "5 alpha reductase 2 deficiency".    Interestingly, the list of "intersex" conditions does not list a prevalence, but the other comments about it being a recessive genetic trait, and actual names given to the condition in the Dominican Republic, Turkey, and Papua New Guinea suggest that what's going on is good old fashioned inbreeding--cue "Dueling Banjos", I guess.  

And knowing a touch about what leads to inbreeding--poverty and social isolation--one would infer that as sports opportunities (and wealth) increase, the conditions for this syndrome will decrease.  So for the sake of young people being born, the world's increasing wealth is a boon.

But if we assume that about one in one hundred thousand live births has this, we would infer that worldwide, there are about 40,000 people with this condition, of which about 5000 are of the ages that would translate to Olympic athletic opportunities.  If we assume that 90% of those afflicted never get a chance to participate, we find that maybe 500 would be athletes, of which maybe 2% (two sigma) would be at the level of "good high school male".  This seems to correspond reasonably well with the few athletes that are clearly "male presenting sort of as female".

The thing that scares me, though, is the thought that some people with this disorder would find themselves at the three sigma (0.135%, D1 athlete) or four sigma (Olympic men's level, 0.003%) ability level, and in any sport that involves contact or combat, the size and strength disparity would go from "dangerous" to "lethal"--think Imane Khalif being replaced by Canelo Alvarez.

As far as I can tell, there is a marginal chance of an occasional D1 level athlete, but the Olympic man presenting as sort of female will be a black swan event unless there is a large mutation that greatly increases the number of people with this disorder, or something like that.  

So I would infer that the big hazard in Olympic sport is not intersex conditions, but rather "trans" athletes using the maximum allowance available for available testosterone.  That said, I still think it's time to bring back the cheek swab and take drug testing seriously.  Crazier things have happened than an XY 5 alpha reductase 2 deficient Canelo Alvarez, after all, and we shouldn't be expecting women athletes to die because we're unwilling to act on the fact that XY is, statistically speaking, bigger and stronger than XX.