Dave Cameron has a little bit to say about dismissing three weeks of evidence based on small sample size. His argument is, essentially, that many in the sabermetrics community tend to cling to the SSS tennent so tightly that they will resist anything that may run counter to their pre-season predictions entirely, when some of the information is actually pretty useful.

I think he has a pretty good point, and uses Victor Martinez and updated ZiPS projections to help make it.

And, he caught the eye of Rob Neyer, who linked to his FG post and summed up the whole idea pretty well, I think:

"Every day gives us numbers that mean something, and every additional day gives us more meaning. And I think that's especially true when you're talking about a team that's made big changes, or a player like Martinez who's coming off an obviously aberrant season. Everything counts."