They aren't wrong because the science proves them wrong, though- they're wrong because they happen to not be right. Sometimes, the scientific or quasi-scientific institutions we have just aren't sufficiently uncentivised to align themselves with the truth! Modern Researchers are beholden to grantwriters, and grantwriters are biased in what sort of peojects they'll give money to. Specifically, low prestige problems are underfunded.

An example I heard about was light-boxes for SAD. There were some studies with a few lights that helped a bit; nobody had done studies with a ridiculous number of lights to see if it solved the problem completely; one guy did, and it fixed his wife's SAD. In absence of data, anecdata rules.

Alternatively, entire fields can be compromised when they're politicised. The soviet union's physics was good- the scientific method worked there, and if you wanted your ideas about physics to match up with reality, you could do a lot worse than copying the most popular soviet physics world-model! Just don't do it for biology, or genetics, or sociology. (Actually, you probably shouldn't trust western sociology, either.)

One person shouldn't be able to outperform the scientific institutions of the world, but it can happen when they're unhealthy.