                        2003 Gold Glove Winners (part 1)
                        --------------------------------

By Tom Tippett
December 5, 2003

[This document is presented in two parts because it is too long to display
in the notes viewer in Diamond Mind Baseball version 8.]

Each year, usually in November, Rawlings announces the winners of their annual 
Gold Glove awards, given to the top fielders in each league.  The winners are 
chosen by a vote of the managers and coaches that is taken before the end of the 
regular season.  

How much weight is put on great range versus soft hands or a good arm or the 
ability to turn the double play?  

One hopes that the voters take all of those things into consideration, with the 
proper weight given to each skill.  But we don't know.  The announcement story 
rarely provides more than the basic info -- who won and how often each player 
has taken home the award.  We're never given any proof that the best man won. 

In contrast, when we're debating the MVP or Cy Young winner, nobody's at a loss 
for words ... my guy deserves the MVP because he nearly won the Triple Crown ... 
no, that's not right, you've got to give it to the man with the 11 game-winning 
hits in the second half ... a 2.20 ERA is worth more than 20 wins because, after 
all, the pitcher doesn't control how much run support he gets ... no, those 55 
saves are far more valuable, because the game is always over as soon as he takes 
the hill, and everybody on both teams knows it.

Not so for the Gold Gloves.  No statistics, no debate, no analysis.  Nothing.

A few years ago, we began trying to fill this void with our own analysis of the 
Gold Glove selections, and we've been at it ever since.  Writing this article is 
a natural extension of the work we do each winter (and have done since 1986) to 
develop fielding ratings for the annual Diamond Mind Baseball season disk.

That work involves looking at defensive performance from many angles in our 
attempt to form the clearest possible picture of the contribution made by each 
player to his team's defensive effort:

- we evaluate team defense using statistics such as the percentage of grounders 
and fly balls turned into outs

- we look at range factors, which are assists and/or putouts per nine defensive 
innings, keeping in mind that range factors can be severely biased by the nature 
of a team's pitching staff:  the left/right mix, strikeout rates, and tendency 
to generate ground balls versus fly balls

- using play-by-play data licensed from STATS, Inc., we compute adjusted range 
factors that take these potential biases into account and focus only on those 
putouts and assists that provide the best indication of fielding skill (catching 
a popup on the infield or taking a throw on a force play are examples of plays 
that generate assists and putouts without telling us much about fielding skill)

- using play-by-play data, we divide the field into zones, measure each 
fielder's ability to turn batted balls into outs in each zone, and compute the 
number of plays each player made above or below the norm for his position given 
the mix of balls hit his way; we call this our "net plays" analysis

- we look at the STATS zone rating and our own zone rating to get another look 
at individual fielding performance, being careful not to be fooled by zone 
ratings that are significantly affected by error rates (our job is to come up 
with separate measures for range and error rates)

- to assess the interaction between neighboring fielders, such as a third 
baseman cutting off grounders that might otherwise be handled by the shortstop, 
we examine the number of plays made by each fielder and by the team in the zones 
where the responsibility overlaps

- we measure the percentage of batted balls turned into outs in home and road 
games to assess how each park might be influencing our measures of team and 
individual defense

- we use play-by-play data to measure other skills that are specific to certain 
positions, such as the ability of middle infielders to turn double plays, the 
ability of pitchers and catchers to shut down the running game, and the ability 
of outfielders to prevent runners from taking extra bases on hits and fly balls

- after all of the individual players have been rated using these methods, we 
cross-check them against our team defense measures to make sure they are 
consistent

- in cases where our findings are at odds with a player's reputation, we use the 
video clips on MLB.com to watch a large number of plays involving that fielder

We believe very strongly that it is only through a combination of these methods 
that one can accurately evaluate defensive performance.  (For a more detailed 
description of this approach, see the Evaluating Defense article on our web 
site.)

Do the Gold Glove voters have this information at their disposal when making 
their selections?  It's doubtful.  More likely, their votes are based on 
traditional fielding statistics, reputations, and appearances.  That's not 
necessarily a bad thing.  In a meaningful number of cases each year, our 
analysis concurs with the Gold Glove selections, in part because the best 
fielders are going to look good no matter how you evaluate them. 

But there are some differences, so let's get right to it.  We'll go position by 
position, commenting on the Gold Glove winners (who are listed in the title for 
all positions other than outfield) and other candidates that we believe were 
deserving of serious consideration.  When we're done, we'll recap by comparing 
our Gold Glove choices to the official winners and offer a few comments on other 
players who caught our eye as we did the fielding ratings for our 2003 Season 
Disk. 


Pitchers (Mike Mussina, Mike Hampton)
-------------------------------------

If you're looking for pitchers who fielded their position without making an 
error, the list begins with Derek Lowe (65 error-free chances), Mark Buehrle 
(53), Mike Mussina (49), Brett Tomko (48), Danny Graves (47), Jon Garland (46), 
Cory Lidle (46), and Mark Mulder (45).

If you can forgive an error or two in favor of a guy who makes a lot of plays, 
then your leading candidates are Tim Hudson (2 errors in 76 chances), Roy 
Halladay (1 in 75), Greg Maddux (2 in 73), Carlos Zambrano (4 in 70), Mike 
Hampton (1 in 68), Derek Lowe (0 in 65), Livan Hernandez (1 in 63).

But this approach is a bit simplistic, mainly because a pitcher's own tendency 
to induce ground balls is a huge factor in the number of assists and putouts he 
gets.  Fielding skill helps, of course, but you can really pad your numbers if 
you can get batters to hit it back to you in the first place.  Five of the 
pitchers we've mentioned -- Lowe, Halladay, Hudson, Mulder, and Maddux -- are 
among the top twenty starters in ground-ball percentage.

A different group of pitchers emerges when you consider the relationship of 
plays made to opportunities.  Among the standouts in 2003 were Kenny Rogers (a 
Gold Glover in 2002), Jae Weong Seo, Jon Garland, and Javier Vazquez.  But it's 
hard to judge pitchers on only one season because they typically get dozens of 
chances to make plays, while other fielders get hundreds of opportunities.  

If we extend our review of pitchers who convert a high percentage of chances 
into outs to include the last three years, the list is topped by Rogers, Steve 
Sparks, Graves, Tom Glavine, Kirk Rueter, Livan Hernandez, Vazquez, Randy Wolf, 
Garland, Steve Trachsel, and Mussina.  Buehrle, Hampton, Maddux, and Halladay 
are a little further down this list.

Mussina was a good pick, in my view, because he was in the league's top tier in 
turning batted balls into outs, was third in the league in error-free chances, 
controlled the running game (only 9 steals in 19 attempts), and has done these 
things well enough in the past to show that this was not a fluke.

But he wasn't the BEST pick.  Kenny Rogers made more plays, both in absolute 
terms and relative to the number of balls hit his way, REALLY shut down the 
running game (only 4 stolen bases allowed all year, 3 pickoffs), and tied for 
second in the league (behind Sparks) with 4 double plays.  Yes, he made two 
errors, but that doesn't cancel everything else, and Rogers gets my vote.

Mike Hampton is similar to Mussina in that he's done enough to be considered a 
serious candidate.  Second in the league in total chances, only one error, very 
good in the running game (only 3 steals allowed in 9 attempts), and a good track 
record.  But Hampton's a ground-ball pitcher who creates lots of chances to make 
plays, and he was only a little better than average in converting those chances 
into outs.

Javier Vazquez, on the other hand, is a fly-ball pitcher who still manages to 
accumulate a good number of successful chances each year.  That's because he's 
always at or near the top of our rankings in converting opportunities into outs.  
And he allowed only three steals in five attempts all year.

Danny Graves is another impressive candidate.  Second in the league in error-
free chances handled, among the leaders in converting chances into outs, both 
this year and in recent years.

But my vote goes to Kirk Rueter.  He handled 43 chances without an error in 
2003.  In fact, he hasn't made an error since 1999, successfully completing 209 
plays in the last four years.  Rueter had a hand in 5 double plays, one shy of 
the league lead.  And he continues to be nearly impossible to run on.  He may 
not have the greatest stuff in the league, but he does a lot of other things to 
keep himself in the game.


Catchers (Bengie Molina, Mike Matheny)
--------------------------------------

Ivan Rodriguez owned this award for a long time, but knee problems have taken 
their toll and it's no longer a slam dunk in his favor.  Still, he continues to 
be a top contender.  Opposing base stealers were successful 68% of the time, an 
ordinary figure, but only one other regular catcher was challenged less often, 
so it's clear that I-Rod's gun still has some bullets in it.  But with 8 errors 
and 10 passed balls, I can't make him my choice.

Mike Matheny was the least-challenged catcher in the majors this year, with a 
runner taking off only once every 19.9 innings.  But those runners arrived 
safely 77% of the time, an unusually high percentage with Matheny behind the 
plate.  Still, St. Louis allowed the second-fewest number of steals of any NL 
team, and Matheny caught in 138 of those games without making a single error.  
He was also second (to Brad Ausmus) in the league in fewest passed balls allowed 
among catchers with at least 1000 innings.

Speaking of Ausmus, he's difficult to evaluate because his manager (Jimy 
Williams) has a history of telling his pitchers to forget about the running game 
and concentrate on the hitters.  It wasn't long ago that Ausmus was throwing out 
half the runners who dared challenge him.  This year, it was only 31%, but 
that's quite good on a Williams team.  Plus, Ausmus made only 3 errors, allowed 
only 3 passed balls, and took part in a major-league leading 10 double plays.

Another candidate was Montreal's Brian Schneider, who led the circuit by 
throwing out 47% of enemy base runners and contributed to 9 double plays while 
making only 3 errors and allowing 3 passed balls.  But Schneider started only 95 
games, compared to 129 for Ausmus and 121 for Matheny, and that hurts his case.  

All things considered, my pick is Matheny by a nose over Ausmus and Schneider.

In the AL, Bengie Molina tied for the league lead by nailing 41% of the runners 
who challenged his arm.  His fielding percentage was only a hair above average, 
but he was among the league's best at preventing passed balls.  His biggest 
weakness is fielding bunts and other balls around the plate, a category in which 
he's been well below average for three years. 

Tampa Bay's Toby Hall is an interesting candidate this year.  He's more agile 
around the plate than Molina, and like Molina, Hall wiped out 41% of enemy base-
stealers.  Further, 81 runners challenged Molina in 950 innings behind the 
plate, while only 78 tested Hall in his 1107 innings.  Only Seattle and Chicago 
allowed fewer stolen bases than the D'Rays in 2003.  On the other hand, Hall's 9 
errors and 7 passed balls are unimpressive.

Chicago's Miguel Olivo is much like Hall.  Olivo may have the league's best arm, 
but his 9 errors and 8 passed balls hurt his case, and he started 28 fewer games 
than Hall did.

If Dan Wilson (92 starts) didn't share the position with Ben Davis, he'd get my 
vote.  He was part of the duo that led the league in fewest steals allowed, he 
led the league in fielding percentage (only one error), and shared the lead in 
fewest passed balls allowed among catchers with at least 800 innings.  But it's 
hard to pick a guy who caught only 57% of his team's innings, so I'll concur 
with the voters and give the nod to Molina.


First basemen (John Olerud, Derrek Lee)
---------------------------------------

In the AL, the voters chose John Olerud for the second year in a row.  In my 
view, it should have been a two-horse race between Doug Mientkiewicz and Travis 
Lee, with Mientkiewicz winning by a few lengths and the rest of the field a long 
way back.

But let's see how Olerud and Mientkiewicz compare:

- Olerud started more games at the position, 143 to 133 ... playing time 
matters, but this is not a big difference

- Olerud led the league in assists with 125 ... but we all know that a first 
baseman can pump up his assist totals simply by making the toss to the pitcher 
while others are taking more balls to the bag themselves

- Olerud participated in 126 double plays, second in the league to Carlos 
Delgado (134) ... this is a legitimate plus for Olerud ... it's hard to judge 
1Bs on overall DP totals because they have little to do with most of them, but 
Olerud was also among the league's best at starting double plays, while 
Mientkiewicz was below average this year and in 2002

- Seattle had the league's lowest error total in 2003, and the lowest number of 
throwing errors, so it's tempting to conclude that Olerud saved his fellow 
infielders a lot of errors ... on the other hand, Seattle was only second best 
in the AL, behind Minnesota, in fewest errors by 2B/3B/SS ... unfortunately, 
it's very hard to measure 1Bs in this manner because our play-by-play data tells 
us how many throwing errors were made, but it doesn't tell us how many throwing 
errors would have been made if not for a good play by the first baseman

- Olerud has a very slight edge in fielding percentage, .998 to .997

- Mientkiewicz has a huge edge in range ... he topped our net-plays rankings and 
was first in the majors in STATS zone ratings, despite playing his home games on 
the fast turf in Minnesota ... Olerud, who is now in his mid-30s and doesn't 
move as well as he did in his prime, has been near the league average in range 
the past four years

All things considered, Mientkiewicz's advantage in range is much greater than 
Olerud's in the other areas, and he gets my vote for the third year in a row.

In the NL, Derrek Lee got the nod for the first time.  In my view, Todd Helton 
and Tino Martinez are the only other serious candidates, but I'll focus on Lee 
versus Helton because both started at least 27 more games than Martinez and 
surpassed him in most key measures.  Here's how I see these two:

- Lee has an edge in fielding percentage with a .996 figure that was third in 
the NL among 1Bs with over 1000 innings.  Helton had an off year in this regard, 
finishing with 11 errors to Lee's 5.

- Lee was second in the league in double plays, and while Helton was first, that 
had a lot more to do with all the DP opportunities that arise when playing in a 
high-offense environment like Coors Field ... Lee was quite a bit better than 
Helton at starting DPs on balls hit to the first baseman

- Helton was in a virtual dead heat with Mientkiewicz for the major-league lead 
in range according to our net plays method, and while Lee also showed very good 
to excellent range, Helton had a sizeable lead in this measure ... Lee had the 
edge in the STATS zone rating, but most of that is due to his lower error rate, 
and we've already taken that into account

- Colorado's other infielders made many more errors than did Florida's, both in 
2003 and over the past three years, perhaps indicating that Lee is better at 
handling bad throws

So we have a big edge in range for Helton and advantages for Lee in errors 
by himself and his fellow infielders and in starting double plays.  Add it all 
up and it's too close to call, so I'll take a page from the NFL's instant replay 
system.  If there's no conclusive evidence, you go with the call that was made 
on the field, and that makes Lee my choice.
 

Second basemen (Bret Boone, Luis Castillo)
------------------------------------------

The AL race should have been between Oakland's Mark Ellis and Anaheim's Adam 
Kennedy.  

This was Bret Boone's second Gold Glove, and as was the case the first time, his 
trump card was reliability.  His .990 mark was good enough to share the league 
lead with Kennedy.  Boone was also very good at starting double plays, though 
it's interesting to note that he was below average before he joined Seattle in 
2001, so his teammates may deserve much of the credit for the improvement in 
Boone's numbers.  He was around the league average in making the pivot on 
potential double play balls that were hit to others.  

But Boone's range has never been anything to write home about.  This year, his 
range factor was second-worst in the majors.  It's true that his range factor 
suffered greatly because he played behind a fly-ball staff, but even after 
adjusting for that and other factors (such as strikeout rate and left/right 
mix), Boone is only a little above average.  In fact, he was in the middle of 
the pack in just about every measure of range that we look at.

Kennedy, on the other hand, has been near the top of our range rankings three 
years running.  Like Boone, he was very reliable.  Kennedy was also above 
average in starting double plays, though not as much as Boone.  Kennedy's pivot 
numbers aren't especially good, but it's hard to tell whether that's him or the 
guy feeding him the ball.  Finally, the fact that Kennedy started only 125 games 
at the position is a negative.

Mark Ellis is a very interesting candidate.  Ellis blew away the competition in 
our net plays analysis and the STATS zone rating, and was near the top (but 
behind Kennedy) in adjusted range factor.  It's not unusual for a converted 
shortstop to shine at second, and Ellis put up very good numbers in a half-
season at the position in 2002.  Perhaps because he is a converted shortstop, 
Ellis lags behind his peers in both starting and making the pivot on potential 
double play balls.  His error rate was average.

In my opinion, Ellis's huge advantage in range makes him more worthy than the 
more polished Boone.  So my ballot, if I had one, would have read Ellis first 
and  Kennedy second.

One more thing before I move on to the other league.  ESPN.com's story about the 
Gold Glove selections included this comment by an unnamed AL coach:  "I voted 
for Adam Kennedy because he made some great plays against us and I happened to 
catch Bret when he made a couple of errors."  We have no way of knowing whether 
this is typical of the amount of thought that goes into the voting, but it 
wouldn't surprise me if it is.

In the NL, my choice is Atlanta's Marcus Giles.  Castillo's fielding percentage 
was a little better, but we're only talking about a difference of one error 
every six weeks.  Castillo has always excelled in making the pivot on the double 
play, but Giles isn't too far behind.  Giles topped Castillo in net plays, the 
STATS zone rating, and range factor (though with the help of a ground-ball 
staff).  It's extremely close, but I'll go with Giles.  

By the way, I think Placido Polanco was the best defensive second baseman in the 
league, but he only made 99 starts at the position before moving to third when 
Philly had to get David Bell's bat out of the lineup.  Pokey Reese's injury took 
him out of the running.


Third basemen (Eric Chavez, Scott Rolen)
----------------------------------------

Eric Chavez took home his third Gold Glove, and I have no quarrel with this 
decision.  Chavez led the AL in many categories, including range factor, 
putouts, assists, double plays, and our net plays analysis.  

His standing in the first four of those categories is a bit artificial -- he 
played more innings than anyone but Tony Batista, his staff induces a lot of 
ground balls, and Oakland had by far the highest percentage of innings by lefty 
pitchers in the majors, so Chavez saw a steady stream of right-handed batters 
who tend to pull the ball in his direction.

Chavez is no Brooks Robinson, but he's solidly above average in range, and he's 
reliable (third in the league in fielding percentage, only a hair behind the 
leader), and he did those things almost every day.

My choice last year was Cory Koskie of Minnesota, who had another very good year 
in the field.  He led the league in fielding percentage and was above average in 
range again, but he's my runner-up this time.  Damian Rolls is someone to watch.  
He didn't play enough (68 starts), and may never hit well enough to be a full-
time player, but he looked good in every measure that we use.

Scott Rolen is a perennial standout who has made far more plays relative to the 
norm for his position than any other NL fielder over the past five years.  But 
his performance showed a marked decline in 2003.  His range factor and STATS 
zone rating were slightly below average.  His double-play numbers, normally a 
strength, were down.  In our net plays analysis, we're accustomed to seeing him 
come in at 40 plays above the league, but he was in the middle of the pack in 
2003.

It's possible that injuries are at the root of this decline.  In the 2002 
playoffs, Rolen collided with a baserunner and sprained his shoulder badly 
enough to keep him out of action for the rest of the postseason.  He has a 
history of back problems and missed games in 2003 with stiffness in his neck and 
back and soreness in both shoulders.  

Still, we're puzzled by the sudden drop in his defensive numbers.  Rolen had a 
very good year at the plate, so his ailments couldn't have bothered him too 
much, at least not while he was batting.   

All in all, it appears that Rolen may have gotten this Gold Glove on reputation, 
not performance.  Having said that, who do you give it to?  Nobody else stands 
out.

David Bell showed terrific range again this year, but his anemic bat cost him 
his job, and he started only 81 games at third.  (Some years, it seems as if you 
can win a Gold Glove with your bat.  Bell may have just lost one that way.)

Adrian Beltre showed good range and posted a league-average fielding percentage, 
so he's a possibility, though his home park helps him look good.  Morgan Ensberg 
was pretty good but only played a half a season.  Craig Counsell and Jamey 
Carroll also look good, but they didn't play nearly enough, either.  Aaron Boone 
was traded out of the league.  Vinny Castilla showed good range and was a plus 
on the double play, but made 19 errors.

It comes down to Rolen versus Beltre, and it appears to me that Beltre had a 
slightly better year in 2003, so he's my choice.  I love watching Rolen play 
third, however, so I hope he bounces back in a big way next year.

[Continued in part 2]
