http://vegaswatch.net/2009/09/evaluating-april-mlb-predictions-2009.html

(opens in new window)

Because it's impossible to quantify everything that affects (potentially affects) a team's record over the course of 162 games (especially unknown future events like injuries), predictions made before the season are usually pretty inaccurate - especially when given for more than one team.  I think I picked the Rangers to lose 90 (or more), and there were (in my mind) a lot of issues surrounding this team that concerned me, warranting that prediction.  I was (mostly) wrong, although the team made a concerted effort plunge toward the end.

It's tough to predict with any accuracy.  That said, it's good/enjoyable to look back and see how everyone fared.  Here's an article doing just that.