I liked this article by Baker that went up this morning. There's lots of interesting stuff in there throughout the whole piece. It seems like, as we've put more distance between ourselves and last season, when it comes to points of intra-blogosphere disagreement we've been able to raise the level of dialogue and reduce the amount of snark and dismissiveness of which we were all guilty, and that can only be good for everyone. Because who wants to feud with a beat writer who actually has the stones to ask Raul Ibanez about his defense?
Anyway, Baker raised two particular points that I want to talk about a little bit. The first concerns Raul's ability to produce at the plate while being an everyday player in the field. There are a few different directions you can go with this, but as for the main one, it's true - Raul deserves some credit for being able to play almost every day. Since 2005, he's only missed 16 games. Durability matters, because the more playing time you get out of your starters, the less you have to get out of your inferior backups. It's helpful to have a guy who can play the outfield every day without getting worn down and requiring a weekly breather.
In the end, though, while durability is good, it doesn't really provide you with that much of a boost. Let's take a hypothetical team that has two players: LF A, who's a 3-win player, and LF B, who's a 0.5-win player. LF A is the starter, and LF B is the backup. How does left field's contribution to the team change as you decrease LF A's projected playing time?
LF A, 160 games: +3 wins
LF A, 150 games: +2.8
LF A, 140 games: +2.7
LF A, 130 games: +2.5
LF A, 120 games: +2.4
LF A, 110 games: +2.2
LF A, 100 games: +2.0
For every additional ten games, you're talking about a small fraction of a win.
It's good that Raul is durable. The last step in the WAR calculation is a multiplier for playing time, and the more playing time you expect a guy to get, the higher his WAR ends up being. But, statistically, a guy who plays a full season is only a few runs more valuable than a guy who misses a little time. It's something, but it's only a very minor selling point.
Maybe this isn't what Baker was talking about when he extolled Raul's ability to be an everyday player. That part was a little unclear to me. If he was referring to the fact that Raul can produce at a moderately high level while patrolling the outfield, then that should show up in the numbers, because he's compared against his peers, and if his peers have more of a tendency to wear down, then that'll make them look worse. But Raul doesn't really stand out from the pack in that regard. So I dunno. Maybe I'm missing something.
The second point I wanted to address was Baker's assertion that offensive statistics are much more accurate and precise than defensive statistics. He's right, obviously. Our measures of how a guy performed at the plate right now are nearly perfect, whereas our measures of how the same guy performed in the field leave a little something to be desired. This is why we can, say, refer to a player as having been worth X runs at the plate but worth Y < Z < (Y + 10) runs with the glove. You can only narrow defense down by so much before you run into trouble.
But with rare exception, we can get those ranges down to a reasonable magnitude. We might as well take Raul as an example. UZR, PMR, and +/- combine to project him as a ~ -15 run defensive LF. They're all in pretty good agreement with each other. And if they're calling him -15, then that means we can say with a reasonable degree of confidence that he's really somewhere between -10 and -20. All things considered, that's a pretty narrow range. Over a full season, it's less than one win.
Because of the error inherent in evaluating a guy's defense, front offices should absolutely put more weight on his bat. They'd be stupid not to. But they also need to keep the guy's defense in mind. You can't just pay a player without thinking much about his glove, because that's a legitimate concern that demands a lot of attention. Teams should always be conservative in their defensive evaluation - that'll keep them from overvaluing defense, as Baker warns against - but they still need to consider a player's estimated defensive ability and factor it into the decisions they make, because not doing so undervalues defense, and that's just as big of a problem. Defensive evaluation is still in its infancy, but it's the teams that do it and the teams that understand it that really have an edge.