A CITIZEN'S BLOG

While I was back at my parents home in Downingtown I snared my copies of George Will’s two baseball books, Men At Work and Bunts. I hadn’t read or thought much about either since I began blogging, but I was curious about what I’d think about both books in light of the evolution in my thinking (namely reading Moneyball and watching the 2004 season unfold) this year. I paged through the index in both books and read all of Will’s entries on Bill James (he specifically talked about James in a 1983 column republished in Bunts on pages 54-56), but found myself drawn to Will’s chapter on defense in Men At Work. (Pages 235-292.)

As everyone who follows A Citizen’s Blog knows, I’ve been doing a lot of work of late on the subject of defense and how it is a critical lynchpin in the Phillies success … or lack thereof. This is a subject not much discussed or recognized in baseball: e.g., look in the Inquirer during the season and check out the Phillies season stats. You’ll see the whole smorgasbord of batting stats (home runs, runs, RBIs, doubles, batting average), and the pitching stats (wins, losses, saves, ERA, strikeouts), but when have you ever seen fielding percentage listed? To illustrate how under-rated defense is in baseball, Will quoted the Washington Post’s Tom Boswell in Men At Work as saying that defense is “the cognoscenti corner of baseball, the poorly lighted room in the gallery.” (p. 274) It is a subject Will clearly enjoyed writing about: in contrast to his chapters on managing, hitting and pitching, the chapter on defense is his most aggressive and authoritative. (Though curiously, while Will devotes one of the chapters in Men At Work to the subject, it is the shortest of the four chapters, and there is no entry in the Index of Bunts for “defense”.) Will described Men At Work as an “antiheroic” book, a book that railed against the individual as master of baseball. (Will is no libertarian: his brand of conservatism is oriented towards the importance of communities rather than the autonomy of individuals.) Defense, Will argues, is the most team-like part of the game (p. 240), and Will exaults the primacy of the team over the individual.

I'll return to Men At Work in the moment, but back to the Phillies ... first off, thank goodness for Google: I’ve been doing a little research on fielding win shares on the 'net and here is what I’ve turned up thanks to google's search engine...

I was able to find a site that had the Phillies 2003 Fielding Win Shares: namely THT's Studes outstanding Baseballgraphs.com (Click here for Baseball Graphs 2003 Win Shares page.) Here they are and their variances with each player’s 2004 performance …

2003 Fielding Win Shares (2004 Fielding Win Shares & 2004 increase / decrease are in parentheses):

1B Thome: 2.4 (1.4 / -1.0)
2B Polanco: 5.6 (5.8 / +0.2)
SS Rollins: 7.9 (4.9 / -3.0)
3B Bell: 3.1 (4.4 / +1.3)
RF Abreu: 2.9 (3.8 / +0.9)
LF: Burrell: 2.2 (2.7 / +0.5)
CF Byrd: 3.7 (2.6 / -0.9)
C Lieberthal: 1.2 (3.7 / +2.5)

IF Utley: 1.7 (2.0 / +0.3)
IF Perez: 2.6 (1.3 / -1.3)
IF Houston: 0.5 (n/a)
IF Punto: 1.1 (n/a)
OF Michaels: 0.4 (2.5 / +2.1)
OF Ledee: 1.1 (1.1 / 0
C Pratt: 0.4 (1.3 / +0.9)
Team: 36.8 (39.3 / +2.5)

-One thing that jumped out at me was the remarkable fact that the Phillies entire 2003 roster basically came back in 2004 intact. No trades, no free agent defections, nothing. Nick Punto and Tyler Houston were gone, but nobody important left or was added since 2003. Team stability is a good thing, but as someone commented on A Citizen's Blog a while back, teams need to change and evolve or they suffer the fate of the 2004 Seattle Mariners.

-Total, the Phillies logged 39.3 Win Shares in 2004 with their gloves and 36.8 in 2003.

-If you go by Win Shares, Rollins 2004 was worse (significantly) in the field then his 2003. Here's the way he measured up defensively from 2002-2004 statistically:

2004 / 2003 / 2002
Fpct: .986 / .979 / .980
ZR: .858 / .840 / .861
Errors: 9 / 14 / 14
RF: 4.01 / 4.42 / 4.58
(Range Factor: Put-Outs + Assists / Innings = Range Factor)

(I didn't bother with fielding win shares per 1,000 innings because Rollins played essentially the same number of innings both years: 1,376 in 2004 and 1,357 in 2003.)

Of course, Rollins had a total of 18.5 Win Shares in 2003, so just 10.7 came via his bat. Rollins was much, much better offensively in 2004: 20.8 Win Shares offensively, giving him a total of 25.7 Win Shares, a 7.2 increase over 2003. Rollins doubled his offensively contribution at the plate.

As I said there was remarkable stability from 2003-2004 (and it appears little will happen to team chemistry in 2005 as well), but the 2002 team was a little different. The Phillies old infield of Scott Rolen, Travis Lee, Marlon Anderson and Jimmy Rollins is just a little different these days. For a little comparison, here are the Phillies 2002 Fielding Win Shares, according to the website Baseball Truth.com:

Lee: 2.1
Rollins: 6.5
Anderson: 3.7
Rolen: 3.6
Polanco: 2.4
Burrell: 2.0
Glanville: 2.2
Abreu: 2.2
Lieberthal: 5.7

Pratt: 1.7
Estrada: 0.3
Perez: 2.1
Ledee: 0.9
Byrd: 0.1
Hollins: 0.1
Giambi: 0.4
Michaels: 0.2

-Here's how Rolen and Polanco stacked up against each other defensively:

Rolen: 4.1 fielding win shares per 1,000 innings / .820 ZR
Polanco: 5.0 fielding win shares per 1,000 innings / .834 ZR

-One interesting part of the defense chapter in Men At Work was a chart Will had tabulating the percentages of balls put into play to each portion of the field. I tried to see if I could do the same via ESPN's fielding stats, but I quickly realized that couldn't, in part because ESPN defines "total chances" as handling the ball in any manner on the play and since first basemen handle the ball on 6-3, 5-3 and 4-3 groundouts, they handle a disproportionate amount of chances.

It's a shame, because I'm curious how the ball get distributed at Citizens compared with other MLB parks. Anyway, here is Will's chart:
C: 1%
P: 6%
1B: 10%
2B: 13%
SS: 15%
3B: 12%
LF: 13%
CF: 18%
RF: 12%

I think this highlights an important reason why I think Polanco needs to get re-signed: the middle defense (SS, 2B, CF) is 46% of your total chances. If the ball is put into play every other time it is going to be handled by just three of your nine fielders. So it is important to be strong up the middle.

Rollins: 3.6 FWS per 1,000 / .858 ZR
Utley: 3.9 FWS per 1,000 / .864 ZR
Polanco: 5.6 FWS per 1,000 / .816 ZR
Glanville: 5.1 FWS per 1,000 / .870 ZR
Michaels: 4.1 FWS per 1,000 / .844 ZR
Byrd: 3.5 FWS per 1,000 / .846 ZR

I think the Phillies are basically strong up the middle.

As I said, the problem with using "total chances" is that first basemen handle the ball on routine groundouts, so it makes evaluating the dispersal numbers difficult. Here's what I got:

2004 Total Chances (excluding first basemen):
2B: 857
SS: 664
3B: 505
RF: 354
LF: 315
CF: 431

It was a little easier to evaluate how the ball got dispursed in the outfield:

2004 (%)/ 2003 (%) / 2002 (%) / 2001 (%)
LF: 315 (29%) / 298 (29%) / 302 (31%)/ 303 (27%)
CF: 431 (39%) / 381 (38%) / 370 (38%) / 469 (42%)
RF: 354 (32%) / 333 (33%) / 307 (31%) / 341 (31%)
Total: 1,100 / 1,012 / 979 / 1,113

Conclusions: as I've said defense is an interesting topic, the stats I listed here are mostly presented without conclusions because I am still piecing together what it all means. What I love about defense as a topic is vague it is: you can't easily assign numbers to these things. Subjectivity is what we see with defense, instinct an almost undefinable and impossible to assign a number to. In Men At Work, Will criticized Bill James' Range Factor as part of “attempts to put artificial precision into a subject that will not hold still for such precision. It is an attempt to attach a numerical value to an activity in which the explanatory value of the number is largely vitiated by the number of variables involved.” (p.273) (To prove the point Will puts in a table showing the defensive stats of a 1989 Mets-Phillies game where no players made any assists: all of the outs were line-drives to infielders, flyouts to outfielders or strikeouts.) (Oh, and here is a link to the box score from that game that I found at Retrosheet.org.) Will himself spent a great deal of time in the book trying to use statistics to show what a terrific shortstop Cal Ripken Jr., was, but he's right: precision elludes us when it comes to rating defense.

But it's still fun to try...

(Check back later today for my thoughts on the Mulder deal and the Eagles 12-7 victory over the Cowboys...)