Jason Michaels

Over at Rotoauthority, I posted a comment asking whether or not ABs are the appropriate unit of measurement in determining how a player will do in the future. I posited it may be a season. (Without getting into the operational defintion of a season, let's just accept it is around 550 ABs.)
I have been tossing this thought around my head for a few weeks. What birthed it from my head to the internet was the discussion about Tim's projection for Jason Michaels.

A common rule of thumb for a part-time player who finds himself in a full-time role is to pro-rate the player's ABs over a full season. Using only 2005, Michaels looks like a single digit HR-hitting OF with no speed, a player who does not tend to grab 600 ABs. However, if you add his 2004 and 2005 seasons together, he has 600 ABs - a full season! Using this method, Michaels is a 14 HR hitting OF - a commodity worth low teens in a draft especially if his AVG holds up. Think Mark Kotsay.

Or how about a player who gets 150 ABs for 4 years. Would his 600 ABs make a good sample or would it be more accurate to conclude he got 150 ABs per year because he could not perform over 600 consecutive ABs? Or that the total ABs is an inadequate measure to predict performance?

I do not know, but I am tending towards the opinion that pro-rating ABs on a straight line basis may be wrong and the way to do it would be one that took a diminishing returns perspective.

Daryle Ward strikes me as a good example. The Astros limited his exposure in his first three seasons. He received about 625 ABs and hit 37 HR. Going into his 4th season, expectations were very high based on that three season stretch, which coincidentally total 600 or so ABs.

He hit 12 HR in 453 ABs.

Is Jason Michaels this year's Daryle Ward?