Wednesday, August 08, 2007

the role of defense in babip

it was pointed out in the comments to my last thread that i've done very little to address explicitly the potential role of good defense in suppressing the babip allowed for the 2007 club. of course i've been broadcasting about this anomalous babip since the first month of the season, and warning about how the club might perform if its good fortune in this respect runs out. what if, it was proffered, the anomaly isn't luck at all -- but instead the effect of great defense converting balls in play into outs at accelerated rates?

this is an effect i examined and dismissed very early on, but have made little noise about. so i'll collect my comments from that thread here a a means of explaining my conclusion.

are the cubs a good defensive team?

it has been pointed out that, by several statistical measures, the cubs look to be an excelent defensive ballclub in 2007. but the difficulty with using current statistics to measure this is that virtually all defensive statistics are deeply interrelated in with babip -- luck and good defense are virtually indistinguishable, speaking statistically. zone ratings, for example -- the fate of the balls in any zone is a product of both luck and skill. which is it this season? i don't think there's any metrical way (that i've seen) to tell.

to my way of thinking, the only reliable way to discern excellent defensive players is through long sample size. the poor zone rating of derrek lee at first in 2007 is a spot of evidence demonstrating the vicissitude of defensive metrics over shorter spans, precisely because luck trumps skill.

so when we look at the cubs' defensive lineups, how many longtime and/or expected standout defensive players do we see?

michael barrett caught half the games for this club, and koyie hill many of the rest. of course, catchers have precious little to do with balls in play, so we can ignore them for now.

mark derosa, mike fontenot and ryan theriot have manned second -- none of them are defensive standouts. examine these players' davenport translations and pecota cards. derosa has managed just 4 fielding runs above average in 170 career games at second base. fontenot was a substandard second baseman throughout the last few years in the minors. theriot was a deficient shortstop and at best an average second baseman.

aramis ramirez has played three-quarters of the reps at third. is he defensively good? he never has been. is he suddenly now? it makes me laugh a bit that the same communities that have decried aramis' obvious defensive shortcomings for years has suddenly been converted into believing that he's really good. he plays third because he can do so plausibly if tenuously while he hits, and if he looks competent in the field over any span it is very likely the product of chance.

theriot and cesar izturis have split short. that's very probably a sub-par defensive platoon. izturis won a gold glove once, but following on leg injury problems that began while playing for the los angeles dodgers was at best an average shortstop and probably not even that good. from 2004 to his trade to pittsburgh, izturis generated (-5) fraa in 248 games.

alfonso soriano in left is a strange figure. are people convinced that he is a good outfielder? i think a case can be made for and against -- lots of assists, in part because teams consistently challenge his arm, but many errors for a left fielder. (in fact, soriano has as many errors as adam dunn in left since the beginning of 2006. consider that for a moment.) the balance to me seems to favor adequacy, not excellence.

jacque jones, felix pie and angel pagan have split center three ways. one can make a case for solidity here but hardly excellence. jones has been long considered a slightly-better-than-average outfielder in spite of his errant arm. pie will hopefully become an even better centerfielder than he is today, as he currently is left to run through many of his mistakes. but he was excellent in limited playing time. pagan has been merely average in center.

cliff floyd, jones and matt murton have split right three ways. all three are problematic. floyd can't run, jones can't throw and murton can't do either.

this, on balance, is a club with fielding flaws but one which certainly could be worse. derrek lee is unquestionably a very good first baseman, and makes his questionable infield somewhat better by saving their errant throws. the decency in center mitigates some of the flaws in the corners. the assessment of the club as fairly average overall comports well with the fact that it places 7th in old-fashioned team fielding percentage. further shooting a hole in the notion that the cubs are a really good defensive club is the fact that they are 16th of 16 in assists. the cubs are a low-chances team by virtue of its pitching leading the nl in strikeouts, but the club obtains assists at a league-low rate of 25.4% of chances.

could even a very good defensive club suppress babip significantly?

it should be noted that this is a point quite aside from the argument over whether or not the cubs are any good in the field, which is certainly arguable in its own right. can we find clear examples of really good defensive clubs? and what are their babip allowed trends?

this season's clear defensive standout team is the colorado rockies. they are the nl-low in errors (46) in spite of being 3rd in chances. the also lead the nl in assist-to-chances ratio at 29.5%. they're starting yorvit torreabla, todd helton, kaz matsui, willy taveras and brad hawpe -- all with multiseason histories of being superlative defenders.

their babip against is .298. factoring out coors field by taking only the away split, it's .291.

next best is probably san diego. third-least errors (62) on the most chances, second-highest assist-to-chances ratio (28.3%). the club stars adrian gonzalez, marcus giles, mike cameron and minor league defensive standout kevin kouzmanoff.

team babip allowed: .287. however, they play in petco park, which is as much a babip aid as coors is a babip hindrance. their away split: .297.

this can go on and on. i've looked at several further examples from past seasons -- 2006 houston, 2006 colorado, 2006 san diego, 2005 atlanta, 2005 houston, 2005 saint louis, 2004 dodgers, the 2004 and 2003 phillies, 2003 saint louis -- and can see no measurable ability of obviously good defensive clubs to seriously mitigate babip.

could the effect exist in theory? yes, certainly. does it exist in reality? probably. is it significant enough to be measurable? no -- and certainly isn't significant enough to explain what has happened to the 2007 cubs.

indeed, in this broader analysis, it becomes clearer that in fact the defensive metrics that are commonly used to claim the cubs are a good defensive team -- all of which are closely tied to the fate of balls in play -- are in fact as heavily distorted by a year of very good fortune as pitching runs allowed has been. they are not evidence of defensive quality; they are merely another symptom of the good luck the 2007 cubs have experienced.

some have even deigned to use defensive efficiency -- which is merely one less team pitching babip allowed -- to claim that the cubs are defensively solid. this is a deeply confused assessment of team defense which presumes that babip is predominantly a measure of the efficiency of a team's skill in converting balls in play into outs.

babip is in fact merely a measure of the fate of balls in play. to what extent does that fate have to do with defense? and to what extent do the varying qualities of team defenses explain variations in that fate? those are the more salient points, and the presumed answer (that good defense is very significant on both counts) is almost surely wrong. as we have seen, even very good defenses do very little to significantly suppress babip in contrast to poor ones.

in summary, i hope this better explains this important ancillary point regarding the 2007 cubs' team babip distortion, which has been by far the most notable, radical and important deviation from expected performance in this season. it is perhaps unfortunate that it coincides with the cubs' luck in this regard finally starting to show cracks. in the 21 games going back to july 17, the cub staff has allowed a league-normal .295 babip -- but in that time allowed just 86 runs (4.10 ra/g). the club is 10-11 in that stretch.

but that also includes the more recent sample of the last 13 games -- .326 babip against, .277 batting average allowed, 61 runs allowed (4.69 ra/g) and a 5-8 record. with soriano out for what now appears to be 4-6 weeks, the club is perhaps least able to withstand a bout of genuinely bad luck on balls in play. let's hope it doesn't come.

No comments: