Also, while the numbers correlate fairly well, it really looks like there are just three levels the “market” sees:

1) lowly regarded: $15M

Yes, likely that season carried more weight in contract talk than it should have. On the other hand, though, it was a much better indicator of likely performance than if, say, a 32 year old were to have a “break out” year.

i sure do agree that it seems like teams are not really interested in defense when they evaluate a player, really, unless he is at short or center.

i think that it is true that most teams just look at last year’s stats when they decide to go after a player. look at gary matthews. look at beltre (sorry.) but my astros are going to have exactly one great glove on the field next year – adam everett, who is simply awesome, and one OK glove, willy taveras, who is still trying to understand what a cutoff man is for. Mo Ensberg is very good, but you betcha they will finally dump him off on someone for a LOOGY. can’t never have too many LOOGYs. you see mo is this terrible wussy who won’t swing at ball 4 for some reason like real men do (see willy taveras) so he has to go cuz we can’t have no guys like that on the team.

i think it is paying for home runs and basically ignoring everything else.

The other thing I was thinking is that there’s pretty clearly a large performance gap between most of this year’s position players and the truly elite players (though J.D. Drew is capable of performing at an elite level if his body holds up). The smarter teams have probably been looking at this year’s position free agents and quietly counting the years until Miguel Cabrera is a free agent. Sure, he’ll cost a bijillion dollars, but he’s actually worth a bijillion dollars.

But a player who exceeds their PECOTA projection, or collapses under it, doesn’t mean either that the player broke the system, or that their season was a fluke.

For instance, given three high-risk, high-reward pitching prospects, PECOTA’s going to take a dim view of their future (with, possibly, a great potential return). If one of them has a great season the next year and the other two blow out their shoulders, that doesn’t mean
a) the dim view wasn’t justified for all three
b) that it missed on all three pitchers

The next year, if you ran PECOTA for all three again, the one who took a step forward will look a lot better than his previous forecast +1 year.

PECOTA, and all projection systems to different extents, say “based on a set of data, this is what I think the player will do next year, and beyond that.”

Raul Ibanez, for instance, busted out of his projections. I don’t think that makes PECOTA’s statement “based on my comparison to historical players through baseball, Ibanez’s performance seems likely to tail off” any less valid. We can’t look at projection systems as the voice of destiny.

I think I’ve gone away from my original point now.

I will say, though, that without a survey of PECOTA’s accuracy, the system seems to either nail the projection or miss way high or way low.

Looking at all positions, just being league average offensively seems to mean 1 to 1.5 wins for premium defensive positions like CF,SS,and 2B, roughly 2 to 2.5 wins for the corner positions and surprisingly (to me anyway), 3 wins over replacement for catcher.

So looking at DMZ’s chart above, the going rate for just being average offensively at a corner position could translate into anywhere from $2M to $10M annually. Considering that defense is absent in the correlation above, an average offensive player (whose in the minus defensive runs range could be dramatically overvalued). I guess, duh, but i think the way DMZ is trying to model the FA market is neat.

Extending back into the past would be interesting. For example, taking “overall ability has mattered less than last year’s stat line,” you reach very different conclusions about the accuracy of this hypothesis if you considered two Mariners who signed the same offseason, Sexson and Beltre.