Let me illustrate: We'll assume that trading returns have a binary distribution. Traders win or lose with equal probability. This is not a great model, but it's good for making ballpark estimates, because it overestimates the odds of a track record like Renaissances.
RenTec's Medallion fund has not had a down year in the past 25. The odds of this are at most 1 in 33 million, using our binary model. Survivorship bias does not begin to explain this; there have not been anything resembling 33 million hedge funds over the course of history. I think 30000 hedge funds is a fairly generous estimate.
Exercise: Suppose Medallion's returns are at least a standard deviation above zero, every year. Calculate the odds of this happening.
This is still the wrong math. I'm not qualified to come up with a great model here, but given that the S&P 500 has only lost value in 5 of the last 25 years the chances of having no losses in the last 25 years are a lot greater.
I've been investing for almost a decade, was lucky enough to sit out the worst in 2008, and I haven't had a down year in 10 years myself. This outcome was mostly luck on my part.
Consider the probability of beating the S&P instead. Same math applies. If that doesn't convince you, look up their monthly returns and do the same calculation.
So presumably, there is a 75% chance that the Medallion fund will lose money (hedge fund = zero sum) for the first time ever, during the next few years. I don't think this interpretation of the maths is correct.