I think the key part you missed is that the statistics of this sort are usually quoted net of fees (which makes sense from the perspective of the investor).
So it's evidence that fund managers take an unequal share of their outperformance, not that they don't outperform the market/benchmark (of course, not all do that either).