Andruw

August 28, 2013

One thing I like about Andruw Jones is that you can find him on line with just one word.   Just type "Andruw", and the only thing that comes up is Andruw Jones.   I had a couple of questions in "Hey, Bill" recently about the Hall of Fame stature of a couple of recently departed stars.   Andruw is not actually retired; he is hitting .232 in the Japanese League.   It’s quite a bit like doing dinner theater in Springfield.

                It is my opinion that, while Andruw had a distinguished major league career, he fell clearly and significantly short of the standard of a Hall of Famer.    My standards for the Hall of Fame are:

                1) 300 Career Win Shares, and

                2)  100 more Win Shares than Loss Shares.   

                If a player meets both of those standards, he’s a Hall of Famer.   If he meets neither, he’s not.   If he meets one but not the other, then we’ll look more carefully at the details.   

                Andruw does not meet either standard.  He’s at 247-178. . .70 to 80% of a Hall of Fame career.  Here are his Win Shares and Loss Shares, by season:

 

 

 

 

Batting

Fielding

Total

YEAR

City

Team

Won

Lost

Won

Lost

Won

Lost

Pct

1996

Atlanta

Braves

2

3

1

0

3

3

.433

1997

Atlanta

Braves

8

11

5

1

13

12

.525

1998

Atlanta

Braves

15

10

7

1

22

11

.678

1999

Atlanta

Braves

16

9

6

2

21

11

.660

2000

Atlanta

Braves

19

8

5

2

24

10

.699

     

 

 

   

 

 

 

2001

Atlanta

Braves

13

14

6

2

19

16

.535

2002

Atlanta

Braves

17

7

6

1

23

8

.729

2003

Atlanta

Braves

16

10

4

3

20

13

.608

2004

Atlanta

Braves

13

12

4

3

18

15

.542

2005

Atlanta

Braves

17

9

4

3

22

12

.644

     

 

 

   

 

 

 

2006

Atlanta

Braves

17

7

2

4

19

11

.633

2007

Atlanta

Braves

11

15

5

2

15

17

.474

2008

Los Angeles

Dodgers

1

9

1

1

3

11

.192

2009

Texas

Rangers

6

7

0

2

6

9

.414

2010

Chicago

White Sox

7

6

2

1

9

8

.534

     

 

 

   

 

 

 

2011

New York

Yankees

6

3

1

1

6

3

.652

2012

New York

Yankees

4

6

1

1

5

8

.406

     

 

 

   

 

 

 

     

187

148

60

30

247

178

.582

 

                Obviously, that’s a really good player there, with won-lost contributions for different seasons of 22-11, 21-11, 24-10, 23-8, 20-13 and 22-12.  He fell, by my estimates, about three solid seasons short of Hall of Fame stature.

                The key question for Andruw is how much we should rave about his defense.    One of the people who asked the question which prompted this enquiry, Phil Dellio, put it this way:  

                          I’m having a hard time getting my head around his viability because a) he drops off a cliff once he hits 30, b) his offensive stats are good but not great when era-adjusted, and c) I’m not as confident evaluating defense as I am offense.   The last is my problem, I know.

                No, it’s not.  NONE of us are as confident evaluating defense as we are offense, and none of us should be.   Batting stats were well enough designed in the 1870s and 1880s that they (batting stats, collectively) were able to grow and develop, over the decades, to create a complex and nuanced portrait of a hitter’s skills.    Fielding stats were very poorly designed, in the game’s formative years, so that over the decades they didn’t evolve or develop at all.   The fielding stats that existed in 1980 were almost exactly the same as they were in 1880.

                Since 1980 we have worked very hard to improve our understanding of fielding stats and to create better stats, but we’re still trying to catch up.   It is a simple reality that we do not understand the role of a fielder in run prevention with anything like the same level of sophistication that we understand the role of a hitter in run creation.   No one does, even the experts in the field.

                I am quite confident that the Win Shares/Loss Shares system evaluates good fielders as good fielders and poor fielders as poor fielders more than 95% of the time.  The issue is scale.   I credit Andruw with a won-lost contribution, as a fielder of 60-30, which breaks down, incidentally, as 35-9 through 2002, but 25-22 after 2002.     Which is common; almost all defensive players are much better as young players than as mature players.   This is one of the ways that we know our system works.

                Anyway, the issue is scale.  We evaluate Andruw as a very good defensive player, with a defensive winning percentage of .667.    But there are people who would argue that this number does not give him adequate credit for his defensive performance.   There are people who would argue this from an analytical standpoint, based on their reading of his defensive numbers, and there are people who would argue this from a scouting/observational standpoint, based on watching him play.    "Your system doesn’t give enough credit to a great defensive player," they would argue.    My system believes that he was about 50 runs better than an average defensive player, over the course of his career.   An analyst might argue that that’s not enough runs.   An observer might argue that we have given Andruw 335 Win Shares and Loss Shares for his batting, only 90 for his glove, and that defense is simply more important than that.

                I have done the best I could to place Jones in the "right" position as a defensive player, and, as I said, I have confidence that my system is generally right on a certain level.   But I can’t prove that those who would give more weight to his defense are absolutely wrong, just as I don’t believe that they can prove that they’re right.   But based on my understanding of the record, Jones’ fielding

                a)  Was only truly outstanding through 2002; after that his Gold Gloves were mostly just reputation, and

                b)   Was not of substantial enough value that we should consider him an all-time great player.

                I would put Andruw in a class with Vada Pinson, Cesar Cedeno, Fred Lynn and perhaps a few others.   Jimmy Wynn and Dale Murphy.   These men, all center fielders, were all tremendous players when they were young—such tremendous players that they didn’t need to develop in order to become Hall of Famers; they merely needed to sustain their level of performance for a reasonably full career.   But, for whatever reason, they weren’t able to do that, and fell short of a Hall of Fame standard. 

                This chart gives Andruw’s Win Shares and Loss Shares with his batting performance, season by season:

YEAR

City

HR

RBI

BA

OBP

SLG

OPS

Won

Lost

Pct

1996

Atlanta

5

13

.217

.265

.443

.709

3

3

.433

1997

Atlanta

18

70

.231

.329

.416

.745

13

12

.525

1998

Atlanta

31

90

.271

.321

.515

.836

22

11

.678

1999

Atlanta

26

84

.275

.365

.483

.848

21

11

.660

2000

Atlanta

36

104

.303

.366

.541

.907

24

10

.699

   

 

 

 

     

 

 

 

2001

Atlanta

34

104

.251

.312

.461

.772

19

16

.535

2002

Atlanta

35

94

.264

.366

.513

.878

23

8

.729

2003

Atlanta

36

116

.277

.338

.513

.851

20

13

.608

2004

Atlanta

29

91

.261

.345

.488

.833

18

15

.542

2005

Atlanta

51

128

.263

.347

.575

.922

22

12

.644

   

 

 

 

     

 

 

 

2006

Atlanta

41

129

.262

.363

.531

.894

19

11

.633

2007

Atlanta

26

94

.222

.311

.413

.724

15

17

.474

2008

Los Angeles

3

14

.158

.256

.249

.505

3

11

.192

2009

Texas

17

43

.214

.323

.459

.782

6

9

.414

2010

Chicago

19

48

.230

.341

.486

.827

9

8

.534

   

 

 

 

     

 

 

 

2011

New York

13

33

.247

.356

.495

.851

6

3

.652

2012

New York

14

34

.197

.294

.408

.701

5

8

.406

               

 

 

 

               

247

178

.582

 

 

Bill

 
 

COMMENTS (37 Comments, most recent shown first)

Jamesammon
Now that Bobby Abreu has probably finished his career, I wonder how he measures up to HOF standards using the same Won-Lost figures? I know he has over 350 career win shares.
2:54 PM Sep 12th
 
therevverend
Brendan Ryan of the Mariners is one of the best defensive shortstops I've ever seen. He lost his job half way through the season. Iwakuma's and Felix's ERA have shot up pretty quick. I'm sure there's some good data there. The M's a couple years ago had a bunch of sabermetric fielding analysis guys in the front office. They've fired everyone but Jack Zduriencik and hired all non-sabermetric god only knows what style analysis guys. So it should be interesting to see what the data says. I know watching Saunders, Ackley, Morse and Ibanez galloping around it's a lot more than one or two plays a week. More like three or four a game.
8:16 PM Aug 30th
 
aagcobb
OT, but I didn't know that Bill had developed a method to calculate loss shares. If you have an article published on it, I'd like to read it.
5:14 PM Aug 30th
 
jemanji
I've always come from Bill's perspective, the argument from sense of proportion ...

1. If +30, +40 runs fielders exist, you would think you could build a +200 runs defense. (Even with their taking a few plays away from each other.) You can't. And nine Babe Ruths will wipe out a team of nine Mark Belangers.

2. I sit in the 3rd deck all the time :- ) and to ME, for what it's worth, the number of MARGINAL balls that outfielders catch doesn't LOOK like two per week. It LOOKS like one every other week.

We're not talking about Andruw Jones versus Justin Smoak. We're talking about him vs. B.J. Upton and Austin Jackson -- average ML centerfielders -- and all those guys are track stars.

In the Olympic 100m, do guys win or lose by 10%?

I think Andruw was super great, but I think the typical ML centerfielder is pretty great too.

.
5:12 PM Aug 30th
 
tangotiger
I agree with Trail's characterization in basketball.

And I also basically agree with his characterization in baseball (or hockey for that matter). Right now, it's giving us a totally different approach from all the other methods, and therefore, serves its purpose in that manner.

At the same time, when I look at the names in the leaders/trailers, nothing really stands out as suggesting it's a bad method. Once we're reasonably happy with the names, then it's the scale that jumps out.

How strong a method it is is still up for more research. For example, did Andruw Jones make 411 more plays than other CF who shared the same pitchers because he's that good, or because he hogged the balls from his corner OF, or because he happened to get more balls hit in his direction? Stuff like that is up for more research.

(Remember, I compare Jones not to his backups, but to ANY CF that his pitchers had at any point in their careers.)
9:50 AM Aug 30th
 
Trailbzr
WOWY (pronounced "wowwie"?) is popular in basketball, where the large number of possessions, small number of players, and fact that even the best players sit out 20% of minutes; allows the use of regression or iteration methods to solve for the value of Player X per minute on the court, without taking into account whether X got statistical credit for doing anything.

Whether it's the most direct route to evaluating baseball defense is another question. Compared to other team sports, baseball offers the huge advantage that the defense has the ball, and there are few contributions a player can make to an out that don't result in some kind of statistical credit. Since very few out plays involve more than three players, there's an awful lot of noise to filter out of a "what happened when he was on the field?" method.

BTW, I was heartened to see a discussion about Clay Dalrymple. I don't remember him as a player, but he was the Orioles broadcast color man in the mid-70s until Brooks Robinson retired.


9:43 AM Aug 30th
 
tangotiger
As an addendum:

www.insidethebook.com/ee/index.php/site/comments/best_worst_wowy_since_1993_through_age_34/

Andruw Jones is +411 over 10.4 full seasons, or around +40 plays per 162G, or +32 runs per 162G.
9:14 AM Aug 30th
 
tangotiger
The study was very basic, and I didn't try to do anything authoratative. Just a first stab kind of thing.

If we look at 1977 Olympic Stadium, I put in one pool all those "long career SS" in one pool, and "everyone else" in the other pool. I looked at the actually innings they played in at 1977 Olympic Stadium, and counted number of runs scored per inning for each pool, and compared them.

I repeated this for every combination of park-year. Because of the way the study is laid out, this means I am only controlling for park-year. I am not controlling for team.

Having said that, I have a different process called "With or Without You" (WOWY) that DOES control for park and pitchers. The basics is laid out here for catchers:
tangotiger.net/catchers.html

I did something similar for SS in a Hardball Times Annual, and luckily Google Books has it available for free here:
www.insidethebook.com/ee/index.php/site/comments/with_or_without_you/

There are two articles: one that expands the catcher one linked above, and another one that applies the same process, but to SS. It's focused on balls in play (and not runs), but this one is more like the one that Bill James was asking for.

9:08 AM Aug 30th
 
chill
[following up my previous comment - this may or may not be relevant] If Tango's study compared Dodgers-with-Russell to Astros-with-Metzger, I'd worry about quality leakage. I'd worry that defensive quality tended to concentrate on the rosters where those all time great shortstops played. That you're less likely to surround Ozzie Smith or Derek Jeter with chumps (including defensive chumps), so that the difference measured between teams with great shortstops and teams without might be the difference between all around good and bad teams, to some extent. And high quality teammates would pollute the data, as they'd lower the runs-scored-against, with the shortstop (in Tom's study) getting the credit. If I misunderstood the study, sorry!
9:48 PM Aug 29th
 
chill
I feel a little silly saying this, but I've read the first bit of Tom's study several times and can't quite get it straight what he did. Tom, did you compare runs allowed by the top shortstops' teams (at Montreal in 1977, say) to runs allowed by those same teams when the main shortstop was out? Or to runs allowed by other teams without a top shortstop? That is, are runs allowed by Dodgers-with-Russell compared to Astros-with-Metzger or to Dodgers-with-Ron-Washington?
9:42 PM Aug 29th
 
tangotiger
First off, I used "50" because someone else used it. I didn't make that specific claim, though my other research suggests that 50 could be that number for the best fielder ever.

Suppose the average CF has 500 plays hit in some large area around CF, and the average CF makes 400 outs. So, in this case, it's .80 outs per play. Is it hard to believe that the best fielder ever could make say .90 outs per play?

If you want to throw out 200 of those 400 outs as "anyone could make those", then we're left with 200 outs on 300 plays (.67 outs per play), and that the best fielder ever would make 250/300 (.83 outs per play). Doesn't seem outlandish to me.

And considering what Dewan is showing for Andrelton Simmons right now, the data can be easily used to show that this is the case.
9:14 PM Aug 29th
 
wovenstrap
Just noticed the Tango comment about the e.g. .90 and .10 outs and so forth, which does make the claim seem more reasonable. But by the same token, every time you adjust it in such a way that the plays are easier to make, then more CFs are going to be capable of them. In the raw sense, the 50 plays a year thing doesn't work, and it may be an indication that it doesn't work even if you slice up the 50 plays.

4:51 PM Aug 29th
 
wovenstrap
So I'll be the first to say it. I doubt it's 50 times a year. Unless I'm misunderstanding the claim Tango made, we're talking about 50 plays per year that Andruw Jones would be the ONLY guy in baseball to make -- that is how I interpret the meaning of "sure-hit to sure-out." Does that play happen once every THREE games in Andruw's prime? (3x50=150) That Andruw makes a play that nobody else in MLB would be able to make? Hell no. Do more than about six guys hit fly balls to center field in a typical game? Aren't most of those fly balls pretty easy to deal with, for a professional center fielder? There aren't enough plays there.

That's why we need better defining. Whether he meant it that way or not, Tango's comment seemed like an attempt to make a high-value claim for Andruw's defense to seem reasonable. It is not reasonable. It's likely that Andruw often made plays that a poor center fielder wouldnt make. It's not likely that he often made plays that nobody else would make.
4:39 PM Aug 29th
 
chuck
Bill, getting off on a tangent here, I know, but regarding Bill Russell's defense, the Monday morning blog from 5/8/09:
(https://www.billjamesonline.com/article1138/?AuthorId=3&Month=5&pg=3) ends with an examination of the Great Dodger infield. Russell's fielding win-loss record is given as 78-29 (.726), which surprised me at how much better Russell was than I remember.
3:04 PM Aug 29th
 
tangotiger
Chitown: I'd suggest checking out some articles on my site: Tangotiger.net

The short of it is that the change in run expectancy from a sure-hit to a sure-out is 0.8 runs.

And when we say that Andrew is worth +50 plays, that's actually +50 sure-hits to sure-outs, or say 100 might-be-hits into sure-outs, or say 150 not-really-hits into sure-outs, etc. We're adding up say all the .90 outs into sure out (+.10) and all the .40 outs into sure outs (+.60), etc. And in the end, you might end up with say +50 outs.


1:25 PM Aug 29th
 
ChitownRon
The post below was in regards to tangotigers last comment
1:21 PM Aug 29th
 
tangotiger
The tough part about determining the "best" fielding SS is that you get into a circular argument if you are not careful. For example, if we decide to first look at runs allowed by team with SS in game, and just use that (i.e., like catcher's ERA), then we'll end up with great SS, good SS, and bad SS who played on great pitching teams.

If I rely on someone else's list, I *could* get that kind of results, because how that person's metric is constructed could be based, in part, on team runs allowed, the very thing we're trying to study.

So, that's why I actually liked my first study (50 SS with most innings played). We don't know WHY they played so much, but for the most part, you have to figure it's because they are likely better than average-fielding SS (as a group). There may be some noise in there (Jeter), but that just makes the results more conservative than they'd otherwise be.

And still, I ended up with some 0.25 runs per game difference.

Anyway, that was just a quick first stab from a few years back. In many respects, this is like the Phil Hughes study that Bill did, trying to focus on team-level results, but only in the specific games involving our interested fielders.

***

It would be interesting for example to repeat this study from a hitting viewpoint. Take the 1B who played the most over the last 50 years, and see how many runs their teams scored. Chances are, you'll have mostly great-hitting 1B, but since we have a great idea of what kind of results we should get, this will at least show what kind of errors or noise we have to contend with.
1:20 PM Aug 29th
 
ChitownRon

Regarding your comment:
[i]The value of getting an out, as opposed to letting it drop in for a hit, is roughly 0.8 runs.

If you think he caught 50 balls that other fielders would have let drop in for a hit, that's 40 runs.[i/]

Your statement above doesn't seem right to me.

40 runs based on a single? I understand the bases may be loaded, or have 1 or 2 other runners on base, or have 1 or 2 outs. What I cant understand the 0.8. number. 40 runs seems a lot to me. Around 25 to 40% maybe.
80%.... I just can't wrap my head around that number.
1:20 PM Aug 29th
 
Magpie
Complete change of subject (feel free to ignore this completely!) but I had always thought Andruw's career made a whole lot more sense if he'd actually been born in 1973 instead of 1977. Which was something I had no reason to believe. I'm glad to see some similarly odd career shapes mentioned. Jones doesn't seem quite so strange to me now.
12:55 PM Aug 29th
 
wovenstrap
By the way, this is article #1000 in the website. Congrats!
12:50 PM Aug 29th
 
wovenstrap
Mazeroski just muddies the discussion here, I think. Maz's status as the "best 2b of all time" or whatever has very little to do with his value on the field, it's like two different planes. Saying "OK well we're going to induct every 'best fielder' at each position" isn't anything that makes any sense to the HOF and nobody actually believes that. Are we going to put the best-fielding LF in the game? Was Brooksie actually the best-fielding third baseman in the game or just the one that had the longest career with a lot of offensive production? Center field is a key defensive position but nobody's really making the case that Devon White should be in the Hall of Fame. And that's correct, he shouldn't be in the Hall of Fame. IMO the current idea that Andruw was "that" good is analogous to the Phil Rizzuto reputation in the 1950s. Nobody really knows how good Andruw was, but you signal yourself as a top-flight baseball observer if you say that he was the best ever. No way, no how would I vote for Andruw for the Hall of Fame.
12:45 PM Aug 29th
 
bjames
Responding to Tango's first post. . .that actually IS exactly the kind of study that I am looking for, in order to be willing to make more aggressive estimates of runs saved. Issues with the study, of course. . .there are issues with all studies. . .but that study designates Bill Russell and Ed Brinkman among the best 9 shortstops of all time, apparently based on using someone else's list. I wouldn't agree that either Russell or Brinkman was an above-average shortstop, let alone an all-time great.
12:39 PM Aug 29th
 
tangotiger
The value of getting an out, as opposed to letting it drop in for a hit, is roughly 0.8 runs.

If you think he caught 50 balls that other fielders would have let drop in for a hit, that's 40 runs.
12:00 PM Aug 29th
 
ChitownRon
I'm not saying Curt Flood was the best of all-time,
but he is the best cf that I have witnessed on a regular basis.

In fairness to Jones, I didn't recall seeing him having to make many
above average plays. I don't discount the reputation, I just didn't see
it. Some of that has to do with Andruws reputation. Players that might go from 1st to 3rd on a single, stop at 2nd base because Andruw is in Centerfield.

His reputation alone was enough to stop players from taking an extra base.
How do you quantify that in a fielding statistic?




11:36 AM Aug 29th
 
mikewright
ooops 17 to 18 runs a year. That's why I don't try the more complicated math.
10:18 AM Aug 29th
 
mikewright
If you will entertain a comment from the eyeball crowd:
Jones is one of the two best centerfielders I've ever seen, the other being Gerald Young. My memory goes back to Maddox. Never saw Mays or Speaker. I think Andruw stayed that way until about 05 or 06.

He played incredibly shallow, got a great jump, never took a wrong step and I can't remember a ball being hit over his head. I remember him running up Tal's Hill in Houston a couple of times in a row in one game. Only time I didn't see him smile.

I think Andruw got to about two balls a week that no other fielder would get to. So let's call it 2.5 over the average CF. That's 62 and a half balls a year.

So does the question become what is the value of those outs? I'm not claiming my intuition is correct, but I'm guessing someone could measure that and then base the runs saved on the value of extra outs.

I would guess at least half of those extra outs would have been singles, so that's 30. Of those lets say 1/3 came with two outs and a third of those came with a runner on second who would definitely score on a single. That's 5 runs saved in a year. So we've got 30 balls in the gap or over his head a year. Runners on base 25 percent of the time? We'll say that saves about 8 runs a year. Add two or three for times when there were more than one runner on. Add, say two runs for outfield assists. That gives us 15 runs a year.

That seems a little low, but I bet it's not 30 runs a year.

My feeling, as a devout Braves fan and a particularly devout Andruw fan is he was better than Lynn, Cedeno, et. al., but not quite a Hall of Famer.
10:16 AM Aug 29th
 
tangotiger
This is not exactly what Bill asked for, but it's another way of looking at it. It's runs allowed in games played by the great SS, compared to runs allowed by the rest of the league (controlling for ballpark).

www.insidethebook.com/ee/index.php/comments/how_many_runs_is_a_good_fielding_ss_worth/


9:03 AM Aug 29th
 
KaiserD2
Anyone, of course, can have their own idea about what makes a great player, or a Hall of Famer, and there may well not be anyone reading these posts and cmments who has a vote. My own view, which I am trying to systematize, is inspired by one of Bill's questions from the Keltner test, namely, "If this guy were the best player on your team, could you win the pennant?" I am researching this quite thoroughly these days, and my TENTATIVE (note caps) standard is 4 seasons with 5 wins above average overall, a figure which is certainly good enough and often is good enough to be the best player on a pennant-winning team. Thus my method pays no explicit attention to lifetime totals or lifetime performance even though those are to some extent correlated with number of outstanding seasons.

By that measure Andruw Jones does not quite qualify but he is so close that I would be reluctant to keep him out. According to baseball-reference.com, whose figures I will eventually adjust in a couple of ways, Jones's five best seasons show 6, 5.4, 5, 4.7 and 4.6 wins above average. Other than people from the last 20 years I'm not sure that I know of anyone with five seasons that good who is not in the Hall of Fame, and no one will be surprised that that is a much better record than that of MANY Hall of Famers and many other people like Minnie Minoso whom people are constantly advocating for.

As for defense, I would like to mention the book Wizardry by Michael Humphreys, which made a very impressive attempt to develop a set of good defensive statistics covering the whole of major league baseball since 1893. He also explicitly compared his method, which relies entirely on stats we have had available all that time and makes adjustments for fly ball and ground ball tendencies of pitching staffs and their handedness, to others, including Bill's, insofar as he could. He rated Jones as the best center fielder of all time and said that he saved over 20 runs a season five times, including seasons of 39 and 48. I certainly agree with Bill, by the way, that fielding statistics aren't as reliable as hitting statistics, but my own suspicion is that this is not necessarily the fault of the measuring system, it's because there is so much year-to-year randomness in where balls are going to fall. At the recent SABR convention some one who counts balls in play for a living told me that a particular player on the team he works for had had an astonishing number of balls hit just within his reach this year. I'm sure that happens to some one every year.

It might be worth mentioning, by the way, that Humphreys found Dale Murphy, Fred Lynn, Amos Otis, Ellis Burks, and Robin Yount to be consistently below average center fielders, and Rick Monday to be even worse.

The Hall of Fame has been so diluted that it's tempting to argue that anyone who would not lower the average quality of people in it deserves to go in, and Andruw Jones certainly meets that standard. By my own personal reckoning he was a much better player than Andre Dawson or Jim Rice.
8:49 AM Aug 29th
 
MWeddell
ErnieSS,

I agree in general with your hypothesis that Hall of Fame voters seem to appreciate a player who excels at one thing -- especially defense -- than one who does fairly well across a broad number of facets of the game.

However, I question whether Bill Mazerowski being inducted into the Hall of Fame illustrates that. As I recall, there were restrictive rules regarding who could be inducted by the veterans committee and that time and, as a player who garnered 40%+ in the BBWAA voting but wasn't in the Hall yet, Maz may have been the only viable candidate for the veterans committee to induct.
7:22 AM Aug 29th
 
sansho1
I've always found the runs saved numbers for outfielders to be out of whack. But they should be less so for CF, as that position represents an efficient defensive market -- good fielders gravitate to CF, and if someone can't cut the mustard in CF they can usually be moved to the less demanding LF or RF. So it should be easier to establish a base line of performance in CF. Nonetheless, for a CF to save 40 runs a year, he'd have to make, what, 50-60 plays over and above an average CF (who is, by definition, a good fielder himself)? Even being cognizant of the inherent observation bias involved (a play not made can register in the brain as a fielding non-event, not a failed catch), it just doesn't seem likely. And I'm a longtime Braves fan who watched Andruw every day.
7:09 AM Aug 29th
 
Steven Goldleaf
It's just tough to account for defense in a way that works out fairly for everyone. If you assume that the game is 50% offense and 50% defense, then you've got to attribute a lot of the latter 50% to pitchers, and then (somehow) apportion the remainder among 8 fielders--how any one fielding position gets more than 4% is almost impossible to justify, because you're forced to give others 1%, and who is sure that shortstop is 4x as important as 1B? When you start positing that a CFer gets 7% and a SS 8%--well, the wheels fall off that wagon pretty fast. So you're left with a pretty low--and dissatisfying --cap on great fielders like Ozzie and Andruw.
6:10 AM Aug 29th
 
guidedogjapan
I wasn't aware dinner theater in Springfield was such a high quality product;-)
1:46 AM Aug 29th
 
ErnieSS
Cool response. I never thought about it that way.

Of course, that opens up a whole other bag of worms...the most obvious one being that Brooks Robinson and Ozzie Smith and their ilk are vastly overrated (because they didn't have nearly as much impact on saving runs as UZR has suggested.)

I love these discussion. Please drive safely.
12:20 AM Aug 29th
 
bjames
I generally refrain from responding to reader's comments about my articles until several days have passed, because I think that it supresses free discussion for me to wade in and sound off on everything that is said. However, since I am driving to California starting tomorrow and may be out of contact, let me offer my thoughts.

Achy Mantis says or suggests that sophisiticated defensive systems. . .and let us say UZR as a shorthand. . ..argue that Andruw is probably the greatest defensive center fielder of all time. The reality is, I believe, that there are no UZR-type defensive metrics for Joe DiMaggio, Willie Mays, Tris Speaker, Richie Ashburn or even Garry Maddox. Saying that he is the greatest of "all time" is really merely saying that he is the best of the last 15 or 20 years, when these systems have come into use.

Yes, I agree that he is; this is not, in fact, a point of conflict between us. My system also would, I'm fairly sure, show Jones to be the best defensive center fielder of the last 20 years or so.

Let me note that I would also say about UZR the same words that I have said about my own analysis: I am quite certain that these methods are generally accurate a high percentage of the time. I am not absolutely sure they are accurate all of the time, but most of the time, they're right. The issue is scale. How many runs did the player save, and how many wins resulted from that?

First, I don't think there IS any consensus among various defensive schemes as to the Scale of defensive contributions, so I don't think that I am abstaining from any consensus on the issue. I don't believe there is any consensus as to how many runs Andruw Jones saved with his glove, even if I stay out of it.

However, I would agree that my estimate of the run impact is smaller than the estimates that many other very good analysts would prefer. I see numbers estimating that Andruw Jones--and other outstanding defenders--save 35, 40, even 45 runs a year. I very much doubt that this is true.

I doubt that this is true, first of all, because it is out of scale with the game. In baseball, if you score 750 runs and give up 600 runs, you finish first; if you score 600 runs and give up 750, you finish last. The entire difference between a first-place and a last-place team is just 300 runs, or 150 each way. I don't see how anyone can sustain an argument that it is reasonably probable that the defense of a center fielder is responsble for 10 or 15% of those runs.

Second, I'm not aware that anyone has ever even made an effort to prove that those numbers are reasonable. I'm not aware of a great deal that goes on in sabermetrics, and it is entirely possible that someone HAS made such an effort, but I've merely missed it. But I haven't seen it.

What I mean is this. Yes, I get the fact that your calculation is that this player's defensive plays had a net value of, let us say, 40 runs in a season. That's not proof.

What would be proof? If you took a field of players (a group of players) who had a defensive value of 40 runs and who got hurt and were missing the next season, and their teams allowed 40 additional runs on average, that would certainly be evidence.

If you took some superior defensive players who changed teams, either by trade or by free agency, and you could show that the teams that GAINED these players got 40 runs better, while the teams that lost these players got 40 runs worse, that certainly would be evidence that the estimate was reasonable. If you did ENOUGH studies of this nature, that would be proof. I haven't seen that kind of proof.

Back in the cradle of sabermetrics, back in the 1970s and 1980s, we had to prove that our runs created/run impact measures were reasonable, and we did studies exactly like that. The Tigers lost Hank Greenberg suddenly one season to injury and another season to the war. The A's sold Jimmie Foxx in mid-career. The Yankees lost Lou Gehrig suddenly. What happened to these teams, when they lost impact bats of that scale? Was the number of runs that these teams LOST due to the absence of the player consistent with our estimates of how many runs they SHOULD have lost?

Those were crude studies. All of sabermetrics in those days was very crude. But we accepted the responsibility to prove that what we were saying was true, as best we were able to.

To the best of my knowledge there have been no parallel studies about fielding (a), and (b) I don't believe there ever will be, because I don't believe that it is true. I don't believe that Andruw Jones--or any other fielder, with perhaps a couple of exceptions--actually had that kind of impact. I don't believe that anyone could demonstrate that he did because, simply stated, he didn't. That's my belief.
12:06 AM Aug 29th
 
myachimantis
Well, it's important to note that if you trust the various +/- systems, Andruw did have a HOF caliber career and is probably the greatest defensive CF of all time. UZR views him as a great defender through 2007.
10:12 PM Aug 28th
 
ErnieSS
Bill, I love your articles, and would like your take on this concept: that the HOF has historically given extra consideration to a player who is widely considered the greatest FIELDER ever at his position.

I assume that this concept must have been the reason that Mazeroski is in Cooperstown, and the reason that Ozzie was a first-ballot, "92% of the vote" guy despite his hitting woes. Wasn't Andruw, in his early days, considered off-the-chart, Mays/Devon White/Curt Flood spectacular at CF? Or was that famous Braves staff an extreme fly-ball staff, which radically padded Andruw's numbers?

I always thought his defense was the reason Andruw was considered a future HOF guy during his career (although that hasn't helped Keith Hernandez).
6:41 PM Aug 28th
 
ErnieSS
Interesting that, according to the Win Shares, Andruw never was remotely close to an MVP season (not even in '05 when he hit 51 HR and came a very close 2nd in the voting).
6:30 PM Aug 28th
 
 
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy