But I doubt there's such a thing as 20% that's going to keep going over anyone's lifetime. Not by means other than criminal.
These figures are net. And for what it's worth, I am in the industry and I don't know any professional who thinks there's obvious fraud going on. It's possible, but you make it seem like there's a consensus that it's illegitimate when you say "most sane people." Frankly it's the other way around.
People who think the returns are fraudulent tend to be outside the industry and thoroughly unacquainted with what quantiles of returns are rare versus implausible. They usually hand wave a misinterpretation of Buffett's famous bet against hedge funds and Fama's (strong) Efficient Market Hypothesis.
I find it highly unlikely that it is about a winning scheme. In my limited understanding of these types of things, it is more likely about a succession of many winning schemes, because these things tend to stop working after a while so you have to find the next inefficiency to exploit.
Both of them have been underperforming in the last decade or so. Medallion is still radically outperforming, however.
Based on what I see from quant prop trader friends of mine in the Chicago area, if Medallion was an order of magnitude smaller they could probably juice their returns up to 150-300% pretty reliably.