The reverse is true, too: Want better performance than people of the same age? Have had more training than them.

I suspect this applies especially to language or logic related tasks.

We will likely never find out, variance seems to be extremely high, while the actual differences are very low.

There are actual studies. Lots of professional athletes were in the older part of their cohort when they were younger. So the effect Terr_ described seems to outweigh the one you describe.