What's interesting is that I'm seeing my usual analytic methods to make plays starts to fall apart at some point mid-summer (August this year) so I have to change tactics and modify how I make selections.
For example, early in the season I focus on 30 day ERA (after a few weeks of play) for starters and team WRC and AVG to narrow my selections and then do some more analysis to end up with a card. That process usually stops working as well and now I've reverted to using more season ERA data, team trends, etc....
Its like you have to move your process as the season progresses.