New guy here... just trying to figure things out.  Just curious if people give much credence to past stats when evaluating players or is it almost exclusively done by the ratings system?  

There always seems to be a few players who play considerably over/under what you think they should based on ratings.    If a guy has a couple of seasons under his belt and is doing something I don't think he should be, I usually assume I'm not looking at something properly.    It's unusual to look at a guy with 1500 AB hitting .300 and continue to think "That dude is a .240 hitter."

As an aside, sometimes it's the park, sometimes the competition in the world isn't so good.  But, for the most part, a .300 hitter after 1500 AB is a .300 hitter.

I look at ratings first, then stats... but what I'm looking for is whether the stats correlate to what I'd expect, or not. If they don't, I try to ask why and learn from it. Like Mike said, it could be circumstantial. But if a player seems to perform "below" or "above" his ratings, I look at the ratings, how they go together, what might be related... then I figure out if it's information I can use going forward, to steer me away or toward similar players.

As far as stats go, if you're looking for ML guys look at ML stats.  I've had plenty of players who kill MiL, but can't hack facing ML players every day.  Tough to argue with past stats unless they've played in extreme parks like Colorado or Tacoma.

Especially for pitchers, even a long record of major league stats can be deceptive, because of the defenses behind the pitcher and the usage strategies. Hitters' stats are somewhat more reliable, but even then rest and platooning patterns can pretty significantly affect stats.

My rule of thumb....

80 / 20 = ratings / stats.