There are several points that one can raise issues here. One is that you assume traders "going belly up" with a 50% mortality rate based on a "rude and crude" linear function. The 50% mortalility rate in itself and the linear function for mortality assumption are not backed by any factual data. In fact, if you have studied mortality functions they are by no means and by any stretch of the imagination linear functions.
Most importantly, it does not address the main question as to how many PhDs are in fact superior traders/investors than the overall population.
On a different note one can also consider what are the probabilities in posting 12 profitable years in a row. Pull out a binomial distribution table for a probability of 50% (or as a matter of fact for any other probability value of an event) and determine what are the odds of posting 12 wins in a row. Not very likely.
But on a different note...even assuming that your assumptions are correct (which seems unlikely) 49 out of 100,000 represents obtaining an event by random chance of less than 0.5% which puts you beyond a 99% confidence interval. If you are a quant in Wall Street, I can comfortably say that they would take these odds any day.
So much for PhDs...
......
