[Drstrangelove]

Regression to the mean.

Given dozens of rookies, one might be expected with random luck to pitch well above average at first, while another might be expected to pitch horribly. Luck in baseball is pervasive. A line drive is hit directly at some one, a squib becomes a hit, a fly ball is caught by the wind, or is carried out, a runner takes off late, or a fielder throws to the wrong base, a hitter just misses on a mistake pitch down the middle, or a hitter swings late and bounces a great pitch through a hole in the infield.

Voila!

Then both of them experience pitching with average luck, (or perchance the luck reverses) and we say:

1) the league caught up with him, or
2) he's figuring it all out

IRL, of course, there are scouts and people learn tendencies and abilities, but in a simulation, the code doesn't need that at all, since human expectations will generate sufficient doubt and optimism, to the generate same feelings.