July 21, 2005

Crooked Numbers

There's no 'D' In Cincinnati

by James Click

Printer-friendly

At Baseball Prospectus, we spend a lot of time reminding ourselves and others not to read too much into small sample sizes. Joe Sheehan said it best early last year: "It's too easy to lose sight of the fact that we just don't have enough information to go on, and to inflate the importance of a frighteningly small sample of games." But we all like talking about baseball, about Brian Roberts hitting .375/.455/.719 in April, about the Yankees' demise, and about the A's struggles, so we talk about them as best we can with constant small-sample-size caveats echoing in the background.

In that vein, it's a good idea to check on how some of the trends we identified early in the season have progressed. For instance, Roberts has hit .329/.393/.520 since May 1, a performance still well above his 90th percentile PECOTA projection, and at the same time far removed from his torrid first month. Similarly, in this space a little over a month ago, the Yankees and Reds were noted as having two of the worst defensive units in the last 30 years, capturing the bottom two spots in this year's Park Adjusted Defensive Efficiency (PADE).

Here's how things shape up only a month later:

Team    Year     DE      LgDE     PF     PADE
----    ----    ----    ----    ------  ------
CHA     2005    .7183   .6945    .9937   3.75
CLE     2005    .7093   .6945    .9993   2.16
TOR     2005    .6995   .6945    .9742   2.04
MIN     2005    .7019   .6945    .9920   1.47
SEA     2005    .7124   .6945   1.0223   1.44
DET     2005    .7036   .6945   1.0016   1.23
PHI     2005    .7005   .6945    .9941   1.16
ANA     2005    .7009   .6945    .9959   1.14
OAK     2005    .7181   .6945   1.0455   1.10
CHN     2005    .7047   .6945   1.0079   1.07
SLN     2005    .7022   .6945   1.0054   0.84
NYN     2005    .6999   .6945    .9997   0.80
PIT     2005    .6959   .6945    .9903   0.70
SFN     2005    .6907   .6945    .9848   0.22
ATL     2005    .6984   .6945   1.0252  -0.69
MIL     2005    .7037   .6945   1.0409  -0.70
BOS     2005    .6887   .6945    .9983  -0.75
WAS     2005    .7025   .6945   1.0411  -0.89
HOU     2005    .6975   .6945   1.0291  -1.01
LAN     2005    .6959   .6945   1.0290  -1.23
SDN     2005    .6907   .6945   1.0165  -1.37
TEX     2005    .6845   .6945   1.0022  -1.55
BAL     2005    .6939   .6945   1.0336  -1.74
ARI     2005    .6775   .6945   1.0011  -2.50
NYA     2005    .6758   .6945   1.0013  -2.75
COL     2005    .6659   .6945    .9732  -2.82
TBA     2005    .6803   .6945   1.0218  -3.11
FLO     2005    .6825   .6945   1.0307  -3.20
KCA     2005    .6742   .6945   1.0066  -3.25
CIN     2005    .6718   .6945   1.0327  -4.83

A month ago, the Reds had a PADE of -5.22; while their current -4.83 would still place them third-worst since 1972, the nearly 40-point jump in a month is impressive. The reasons for the improvement are threefold:

• PADE is measured against league average defensive efficiency (DE) and the league DE has come down from .6955 to .6945 in the last month.

• Great American Ball Park has always had a favorable park factor for defenses, but it's come down from 1.0381--meaning an average defensive team would turn 3.81% more balls in play into outs in that park--to 1.0327, a significant drop considering the park factors are three-year factors.

• The Reds' raw DE has improved from .6712 to .6718, a small difference, but considering the rest of the league has decreased by .0010, their gain is somewhat more impressive.

The bigger story is the Yankees, who have jumped from 29th to 25th, passing the Royals, Marlins, Devil Rays and Rockies and taking themselves completely out of the running for the Worst Defensive Team Since 1972 award. Like the Reds, the Yankees have benefited from the league DE coming down a little, but unlike the Cincy players, the Yanks' park factor has actually risen slightly, meaning Yankee Stadium is playing slightly easier than it was a month ago.

Without question, though, the reason for the Yanks' climb out of the cellar is due to their improved play on the field, jumping from a raw DE of .6674 to .6758. Much like Roberts' hot April skewing his total stat line a bit, the Yanks managed a historically terrible DE of .6551 in April before improving to .6616 in May, .7050 in June, and .6889 in July. How much of this is due to the new deployment of players in the field and how much is just regression to the mean is anybody's guess, but the fact remains that the Yanks have dramatically improved their defense since a brutal first month.

While we're looking at the mid-season defensive numbers, there are quite a few other interesting points to note:

• The White Sox are nearly as good as the Reds are bad. Leading the league by a wide margin, the Sox have turned nearly four percent more balls in play into outs than a league average team would be expected to in their park. It's been said before and I'll echo it: the White Sox are not winning because of their offense, they're winning in spite of it, and their defensive improvement this season is a huge part of it.

• The Rockies--who usually find themselves gaining quite a bit on the field from raw DE to PADE because of the cavernous dimensions of Coors Field--are still among the league's worst performers with the gloves. As if playing in Coors wasn't hard enough on the Rockies' pitchers, they have to deal with a brutal defensive team behind them as well. It's unquestionably one of the worst pitching situations of all-time. A's fans who hope that Joe Kennedy's transition to a friendlier park and better defense will pay off have reasons to be optimistic.

• Last year's PADE leaders, the Dodgers, have fallen off dramatically, going from a PADE of 1.83 in 2004 to -1.23 so far this year. That turnaround of more than three percent isn't solely responsible for their frustrating performance thus far this season, but it's certainly not helping things. It's even more impressive that they managed this collapse despite jettisoning Shawn Green and his atrocious fielding.

• The top six teams in PADE are all in the AL, and four of those are in the AL Central, the runaway Best Defensive Division in Baseball. For hitters in the AL Central, particularly those on Kansas City, that's 60 to 80 games against some of the best defenses in the league. If things even out next season, hitters on those teams could see a bump in BABIP.

• Unsurprisingly, the NL West comes out as the worst defensive division, edging out the AL East (-1.26) with an average PADE of -1.54. The worst division since the '94 AL West, the NL West leads the AL East in pitcher VORP 287.1 to 281.4, but they trail all four other divisions, three of whom are over 600. It's an amazing gap, no doubt widened by their defensive ineptitude.

• The hardest parks in which to play defense: Coors Field, the Rogers Centre, SBC, PNC and the Metrodome. The easiest: McAfee Coliseum, RFK, Miller Park, Camden Yards and the GABP.

For those of you hoping to peruse these numbers yourselves, we're working to get PADE into the regular stat reports soon. With the awesome new sortable stats here at BP up and running beautifully, hopefully I'll be a little more motivated to get them in there so that the next time you want to know how the Reds' assault on the '86 Mariners and '79 A's is going, it won't take waiting on a column from me to get the information.