if there's any group of baseball fans who appreciate the value of good defense, it's cardinal fans. all of our winning teams of the last 25 years have relied heavily on glovework, especially the classic up-the-middle kind. players who contribute mainly with the glove (ozzie, matheney, pendleton, molina) can be as popular in st louis as .330 hitters and 40 hr men. we applaud not only highlight-reel defensive plays but also more workaday stuff -- a strong peg from the outfield that forces a baserunner to retreat, a crisp pivot on a dp. we appreciate good defense so much that we even applaud good plays by the opposing team that rob cardinal hitters of base knocks.
but when it comes to measuring, statistically, the defensive ability of individual players, i sense that we're like every other fan base -- skeptical. i make the observation after having read two articles yesterday, one by david gassko at hardball times and the other by jon weisman at si.com, about defensive metrics. "Throughout baseball's history, fielding evaluation has been more art than science," weisman observes; the naked eye is considered the truest barometer of a player's ability afield, and attempts to quantify it (even trusty old fielding average) are waved aside.
we now have a much finer quantiative grasp of defense than ever before, and more sophisticated systems that can express fielding ability in terms of runs saved or runs allowed. yet fans -- even cardinal fans -- are slow to accept the conclusions proffered. we may embrace the hard numbers when they confirm our prejudices, but when they are in contradiction, we're more likely to trust our prejudices than the numbers. and even when they're in agreement, we tend not to accept the precise number a measuring system spits out. if UZR says jim edmonds saved 22 runs last year, we don't really believe that the cardinals would have allowed exactly 22 more runs if they'd had an average defender in centerfield; we still consider it an eyeballing exercise, ie "edmonds was at the top of the league."
and when evaluating trades or free agent signings, we tend to ignore this stuff altogether. we only consider a player's defense if he's a middle infielder or if, at other positions, he has a reputation for particularly good or bad d. but most of the time we just look at the guy's batting line and figure one fielder is about as good as another.
my question is: why are we fans so reluctant to "go numerical" vis-vis defense when we cite number after number when discussing batters or pitchers? the most commonly cited reasons are that we perceive defense as a) a team domain, and b) inextricably linked to pitching. the batter is alone at the plate, and the pitcher alone on the mound, but the fielder shares his space with eight teammates. and even though the more advanced metrics go a long way toward removing contextual factors from defensive performance, we still resist viewing the fielder in isolation.
here are some other possible explanations for our misgivings about defensive metrics:
-
lack of benchmarks. ev'yone knows what it means to hit .300, or to hit .225; but if your home-team second baseman is +9 on UZR, what the hell is that?
-
lack of consistency.the metrics can seem wildly inconsistent, both internally and with each other. for instance, UZR had edmonds at -15 runs in 2004 (ie, he cost the team 15 runs with the glove) but +20something in 2005; UZR had orlando cabrera at +14 in 2005, while david pinto's PMR put cabrera at -12. which of these disparate results most closely corresponds to reality -- and how are we supposed to know?
- lack of interest. chicks dig the long ball; nobody digs the ability to cut off gappers.
personally, i don't place full faith in any single defensive metric -- mainly due to lack of consistency. if i want to know about a given player's defense, i'll check a range of metrics and see if all the arrows are pointing in the same direction. if they are, i'll feel comfortable with the generic conclusion that a guy's glove saves (or costs) an unspecified number of runs for his team. but i'm a long way from accepting an argument that this fielder is 10 runs better or worse than that fielder.
curious about how all of you feel.