I know generally nothing about the defensive metrics, but I find this incredibly hard to buy into.

I'm glad someone posted this because no matter how much the numbers make sense, and how intuitive the process is, the idea that Dunn and Morgan have the same value seems so ludicrous to me that I just refuse to believe it. I'll continue to stare at the sun, even after reading the textbooks, asserting that the coming eclipse is because of a mean ol' dragon.

The difference between offense and defense is simple: there is a limit to the amount of outs for a defender to make (meaning, a cap on his defensive production) but there is absolutely no cap on offensive production. I've heard anything from defense being half the value of offense to nearly equal to a quarter of the value and so on. Dunn is a phenomenal offensive player because he's very good at not providing outs--despite the K's. To say that Nyjer Morgan is anywhere near Dunn is almost crazy, but then you remember Dunn is awful defensively and that Morgan is good with the glove...

But to be equal? Nah.

When the Nationals signed Adam Dunn over the winter to a 2 year, $20 million contract, the reaction from the sabermetric community was almost unanimously positive towards the move for Washington.

Funny, I don't remember it quite that way.

Okay, so I went ahead and checked Dunn and Morgan's 2009 WARP... and damn it all, they're both at 3.9.

Shoooooooot.

I'm glad I'm not alone in my skepticism. Whew.

The difference between offense and defense is simple: there is a limit to the amount of outs for a defender to make (meaning, a cap on his defensive production) but there is absolutely no cap on offensive production.

This is the difference between offense and defense? Really?

Functionally, you agree, there's a cap to offensive production as well, right? The hypothetical you're talking about doesn't have an actual impact on how baseball games play out, or on the overall values of the players. It's not like Dunn never makes an out, nor are the rest of the Nationals 1.000 OBP players either.

There are salient differences between offense and defense, but the cap to production doesn't seem to be one of them.

I'd like to tell a scout that Morgan = Dunn and watch his head explode

This is interesting. I have hope this will be an excellent thread.

I think in peoples' minds Dunn is a perennial .410 OBP guy. But he's just not.

Cap on defensive outs: 27
Cap on offensive outs: 4-6

I am never sure whether to trust defensive metrics, but I trust them here. In the admittedly few times I have seen him play, Dunn has played defense like a fire hydrant with palsy. While LF doesn't demand much, he's got to be giving up a bunch of his bat with that glove. Now, his bat is great. I do not know if these two are equal-value players long-term. But I absolutely believe they're closer than they look.

The real reason this move was panned, of course, was Milledge's age and potential. So the article is misdirection at best. Still, I hate watching terrible defenders, so I am partial to the basic thesis.

Of course, somebody had to go to AAA to make room for Morgan.

I was afraid this was what would happen. Hopefully, another trade will see him brought back.

Fangraphs - Cameron : Morgan = Dunn

I just saw this on the sidebar and thought it was an SAT style question. I couldn't figure out what "Fangraphs minus Cameron is to Morgan equal to Dunn" was trying to say. That's a minute of my life I'll never get back.

I agree that it does seem odd that total scrubs are good players. I'm still having trouble with the idea that Craig Counsell has been worth $45 million over the last 8 seasons.

Your ideas are intriguing to me and I wish to subscribe to your newsletter.

Dunn has played defense like a fire hydrant with palsy.

thanks to interleague play, I got to finally witness it.

Analysts I have quite a bit of respect for, like Keith Law, Dan Szymborski, and our own R.J. Anderson, hailed this as an easy win for the Pirates, as none of them see much value in Morgan.

This isn't true. I find this a poor trade because Morgan's value is concentrated at a time where it's not particularly valuable to the Nats and there's little long-term upside to Morgan.

And yes, Morgan's as valuable as Dunn, when Morgan's saving more runs on defense than anyone in the history of baseball and having the best offensive season of his career.

Seriously, raise your hand if you think Morgan's going to continue to prevent 23 runs a season more than the average leftfielder. Or 35 runs a season more than the average centerfielder.

In all likelihood, he's a nearly-30 very good defensive centerfielder who is likely to have trouble OPSing in the .650 range on a regular basis. That has value, as I noted.

But where's the upside needed for a team that's 22-53? The Nats don't need average players, the Nats need players that have a chance to be good. If a 32-year-old Morgan declines defensively to simply being a solid +5 player on defense, which is extremely likely given the sample sizes of his career, then he's a 5th outfielder, not a league-average starter.

If I'm down by 6 points on my own 20-yard line in a football game with 2 seconds to play, I go for a Hail Mary pass, even though the expected yardage "on average" will be less than a simple quick slant to a wide receiver. Why? Because, in the position I'm in, one needs to take the high-risk, high-payout play. If I make the Nyjer Morgan play, I'm sitting on the 32-yard line and the game is over, but if I make the Lastings Milledge play, I may throw a pick, but who cares? The game's over anyway. But now at least I have a chance for the high-value result.

Seriously, raise your hand if you think Morgan's going to continue to prevent 23 runs a season more than the average leftfielder. Or 35 runs a season more than the average centerfielder.

In all likelihood, he's a nearly-30 very good defensive centerfielder who is likely to have trouble OPSing in the .650 range on a regular basis. That has value, as I noted.

That's the real problem with Dave's argument. Dunn's value is projectable, Morgan's just isn't.

Is anyone really any better than a true-talent +10 CF? It seems hard to believe, given the average skill of an MLB CF, that anyone this side of prime Willie Mays/Joe DiMaggio even sniff +15 consistently.

Is anyone really any better than a true-talent +10 CF?

Matt Wieters?

Matt Wieters?

His 247/297/388 line hasn't killed that meme?

But where's the upside needed for a team that's 22-53? The Nats don't need average players, the Nats need players that have a chance to be good. If a 32-year-old Morgan declines defensively to simply being a solid +5 player on defense, which is extremely likely given the sample sizes of his career, then he's a 5th outfielder, not a league-average starter.

You're right in that there's no upside to Morgan. He's as good as he's ever going to get.

But this is really as much a move for the near-term development of the pitching staff as anything.

As I said elsewhere, I am skeptical of the UZR rating for Morgan in CF. He has only played about 50 games in CF over his career, and that isn't a very big sample for defensive metrics. He has played mostly LF, and it's plausible that he is as good as UZR shows him in LF. But his UZR/150 in CF is more than 10 runs better than LF UZR (career), which seems implausible to me. Most players will have better range metrics in LF than CF. Cameron recognizes that the UZR result seems too high and reduces the UZR to +10 for Morgan, but how do we know if that is too much or not enough, given the small sample size in CF?

Whether Dunn is a good acquistion for a team partially depends on the positions you have available and the other defenders which surround him. I always felt that putting Dunn is RF, as the Nationals did on the occasions I watched them, is a bad idea, and it's made even worse if the other two outfielders also have terrible range. If a first base position was available, Dunn might make more sense. Conversely, Morgan makes sense for the Nationals because they desperately need a CFer. If they were forced to play Morgan at 1st base, his value would be minimal.

This isn't true. I find this a poor trade because Morgan's value is concentrated at a time where it's not particularly valuable to the Nats and there's little long-term upside to Morgan.

Both things are entirely true about Dunn as well. Far more so, in fact, when you consider that Dunn is costing the Nationals $10 million per year in salary that could have easily been used to fund the restoration of the farm system.

And yes, Morgan's as valuable as Dunn, when Morgan's saving more runs on defense than anyone in the history of baseball and having the best offensive season of his career.

Actually, when he's doing those things, he's twice as valuable as Dunn. Morgan's posted a +4.2 WAR as a major leaguer in basically one full season's worth of playing time. Dunn hasn't posted a +4 win season since 2004, the last year he didn't totally suck at defense. The equality statement was based on an updated ZIPS projection that has Morgan as a .307 wOBA hitter going forward and a +10 defensive center fielder. It's not based on current numbers.

Seriously, raise your hand if you think Morgan's going to continue to prevent 23 runs a season more than the average leftfielder. Or 35 runs a season more than the average centerfielder.

Not raising my hand. I stated in the article that I think Morgan is something like a +10 to +15 defensive center fielder.

In all likelihood, he's a nearly-30 very good defensive centerfielder who is likely to have trouble OPSing in the .650 range on a regular basis

ZIPS projects a .670 OPS going forward that is heavily slanted towards OBP, the more valuable of the two components. The idea that Morgan is a terrible, god-awful hitter is just a myth.

But where's the upside needed for a team that's 22-53?

The upside is in actually developing a competent pitching staff, rather than forcing them all to look like failures because of the ineptitude of their teammates. You can't develop pitchers with the worst defense in baseball behind them and expect it to work out very well. For a team with a rotation full of young arms, giving them some confidence that it's okay to throw strikes without cringing at what might happen if the batter makes contact is a good thing.

I've heard anything from defense being half the value of offense to nearly equal to a quarter of the value and so on.

It depends on the position, but typically the range of performance on defense is about half of the range of performance on offense.

-- MWE

But where's the upside needed for a team that's 22-53? The Nats don't need average players, the Nats need players that have a chance to be good.

they need both, and they have Morgan under control for a few years.

How many seasons are they supposed to punt?

Is anyone really any better than a true-talent +10 CF? It seems hard to believe, given the average skill of an MLB CF, that anyone this side of prime Willie Mays/Joe DiMaggio even sniff +15 consistently.

Erstad put up numbers in the +30 range for a few years. Even with regression, he's gotta be coming in at +15 or so.

I think Morgan is something like a +10 to +15 defensive center fielder.

Echoing snapper, really? Those guys are few and far between.

Both things are entirely true about Dunn as well. Far more so, in fact, when you consider that Dunn is costing the Nationals $10 million per year in salary that could have easily been used to fund the restoration of the farm system.

I wouldn't have traded Dunn for Milledge and Hanrahan if I were the Nationals, as well.


ZIPS projects a .670 OPS going forward that is heavily slanted towards OBP, the more valuable of the two components. The idea that Morgan is a terrible, god-awful hitter is just a myth.

True, but I'm allowed to disagree with ZiPS, am I not? He's absolutely helpless against anything with any break.


The upside is in actually developing a competent pitching staff, rather than forcing them all to look like failures because of the ineptitude of their teammates. You can't develop pitchers with the worst defense in baseball behind them and expect it to work out very well. For a team with a rotation full of young arms, giving them some confidence that it's okay to throw strikes without cringing at what might happen if the batter makes contact is a good thing.

The difference between a +15 CF and a 0 CF is a tenth of a run a game. Pitchers also like not having 7 IP/2 ER starts that result in losses, too.

How many seasons are they supposed to punt?

More than the number of seasons Morgan is going to be good.

Dave raises an interesting idea in that their pitchers might develop better with better defense behind them. I have no idea if that's true, but it's plausible. Also doesn't seem like the hardest thing in the world to test.

Erstad put up numbers in the +30 range for a few years. Even with regression, he's gotta be coming in at +15 or so.

Erstad's pretty unusual in that regard.

However, Morgan's sample size in center is ridiculously small. Morgan has only 51 defensive games at center and 145 defensive games overall in the majors. His centerfield stats, as impressive as they are, are only the equivalent, in predictability of 2-3 weeks of hitting.

The difference between a +15 CF and a 0 CF is a tenth of a run a game.

How many outs is that? How many extra pitches? What about the feedback a pitcher gets. What if he connects a bad outcome (he's not going to really weigh the quality of the fielders in his mind) with a pitch in a certain location or in a specific pattern?

This deal almost ensures the Nick Johnson will be traded before the deadline, with Dunn moving to first base and Dukes returning from Syracuse to start in left. Willie Harris can then resume his role of a utilityman who plays about four times a week (assuming he isn't dealt, too).

I wouldn't have traded Dunn for Milledge and Hanrahan if I were the Nationals, as well.

Okay, great. But given the reaction to the Morgan trade, why not the same anti-Washington rant when they signed Dunn, which was a move that did far more to limit their ability to rebuild and just as little to help them win?

True, but I'm allowed to disagree with ZiPS, am I not? He's absolutely helpless against anything with any break.

Interestingly enough, though, pitchers still throw him fastballs 66% of the time. He is awful against breaking balls, but as long as they keep throwing him fastballs, he'll keep being good enough at the plate.

The difference between a +15 CF and a 0 CF is a tenth of a run a game. Pitchers also like not having 7 IP/2 ER starts that result in losses, too.

If the Nationals had a +0 CF, they might not have made this deal. They don't - they have a brutal outfield of DH's that is easily the worst in the league.

However, Morgan's sample size in center is ridiculously small. Morgan has only 51 defensive games at center and 145 defensive games overall in the majors. His centerfield stats, as impressive as they are, are only the equivalent, in predictability of 2-3 weeks of hitting.


So, what would you project him at in CF, given the sample size and the necessary regression? +5, +7?

BTW, I'm not arguing that Morgan is not a good acquisition for the Nats (they definitely need somebody that can catch the ball), I'm just arguing against the comparison to Dunn.

As someone who crapped all over the Nats for this deal, I'll admit Cameron makes an interesting point. Seattle's having success with the defense-first model this year, and defense is looking like today's "Moneyball" undervalued asset. Morgan's certainly more valuable than his offensive numbers alone would indicate.

The odd thing, though, is that the Nats dumped Ryan Langerhans - a good defensive OF in his own right - the day before giving up Milledge to get Nyjer Morgan. (Langerhans landed, naturally, in Seattle.)

Cameron's analysis is a bit odd in that he's comparing Morgan to Adam Dunn, who has nothing to do with the question of acquiring Morgan. Why not look at the other options Washington had in CF, or compare Morgan to Milledge, factoring in the 5-year age difference? Morgan's defense is likely to decline quickly as his speed diminishes, and at 29, that could start sooner than the Nats would like to think.

Morgan has played a lot more LF than CF, and his LF numbers look a lot more like a +10 CFer than they do a +35 CFer. If you weight and add those is and THEN regress (granted, you'd regress less), you probably get closer to a +5 defensive CFer, maybe +7. I'd have to run the numbers to be sure, and building an entire defensive projection system for this one purpose seems like a bit of a stretch.

However, Morgan's sample size in center is ridiculously small. Morgan has only 51 defensive games at center and 145 defensive games overall in the majors. His centerfield stats, as impressive as they are, are only the equivalent, in predictability of 2-3 weeks of hitting.

Well, sure, but we also have his left field stats, which we can use to infer ability, since the move from a corner OF to CF is one of the easiest to analyze.

I'd say that Morgan's career defensive stats are equivalent to two months of hitting. Which is why I regressed them back so heavily.

And, for those wondering, here's the list of center fielders that we could reasonably presume are in the +10 to +15 true talent range right now.

Carlos Gomez, Franklin Gutierrez, Darin Erstad, Coco Crisp, and maybe Mike Cameron, Carl Crawford, and Rajai Davis (plus pre-injury Endy Chavez). It's definitely select company, but it's not a once in a generation kind of deal. If I was trying to argue that he was +20 to +30 or something, sure. But I'm not.

If you weight and add those is and THEN regress (granted, you'd regress less), you probably get closer to a +5 defensive CFer, maybe +7.

This is what I did.

Cameron's analysis is a bit odd in that he's comparing Morgan to Adam Dunn, who has nothing to do with the question of acquiring Morgan.

Well, one could make the case that if you've installed Adam Dunn in LF that you'd be, looking at things on a day-by-day level, especially interested in a excellent defensive CF who can cover a lot of ground if that's what you think Morgan is.

My quarrel with this is that a team in the Nats' situation should probably be more concerned about upside than about such day-to-day tactical situations. That they are going to finish in last place this year is more or a less a foregone conclusion, unless the Marlins have another fire sale and start a AAA squad for a couple of months.

Matt Wieters?

His 247/297/388 line hasn't killed that meme?

Ah, philosophy: Is there a Matt Wieters meme that even Matt Wieters could not kill?

Why I still take defensive stats with a large grain of salt: Andruw Jones. According to Fangraphs, he was an awesome CF from 2002-07, like +20 or +25 per year. This is UZR with BIS data. If you use STATS UZR instead, he was close to average over that same period.

Given this kind of thing, don't pretend that +10 in CF is a conservative projection for Morgan.

Carlos Gomez, Franklin Gutierrez, Darin Erstad, Coco Crisp, and maybe Mike Cameron, Carl Crawford, and Rajai Davis.

Are you basing that on production or true talent?

Going by FanGraphs UZR
Crisp has some ~+20 years, but he also has a -15 and a couple seasons worth of 0.
Cameron is +5 for his career, and only has one +15 year (last year).
Crawford is +15 in LF, so in CF he'd have to be much worse.
Gutierrez I'd give you.
Gomez and Davis don't have nearly enough sample size, and Erstad doesn't play CF anymore.

I'd guess there are a lot fewer +15 CF's than you surmise, just b/c the average CF is so good.

Given this kind of thing, don't pretend that +10 in CF is a conservative projection for Morgan.

Concur

Okay, here's what I come up with.

Take his UZR/150 for LF, convert to CF using a positional adjustment (10/162*150 = 9.25), and I get 13.25. A 35 UZR/150 for CF. Weight by DG:

(13.25*87+35*51)/(87+51) = 21.29

Regress that to the mean by 2/3, or in other words:

21.29 * 1/3 + 0 * 2/3 = 7.10

What am I mising here?

This is UZR with BIS data. If you use STATS UZR instead, he was close to average over that same period.

BTW, if you missed it, those early year threads on the book blog are well worth reading. The differences between UZR with Stats vs. BIS data on some players (like Andruw) can be astonishing.

Far more so, in fact, when you consider that Dunn is costing the Nationals $10 million per year in salary that could have easily been used to fund the restoration of the farm system.

Funding the farm system and signing free agents aren't mutually exclusive. Or at least they aren't necessarily mutually exclusive, in part because a decent high-profile free agent can pay for himself quickly. That's the whole point of the Value field on FanGraphs' player pages.


Colby Rasmus might be a +15 CF.

[45] Get greenback a prize!

This is a team that's making something like $30 million in profit a year. Signing Dunn shouldn't prevent them from signing int'l players or investing in the farm.

Said another way, the cheapness they've shown there has zippo to do with Dunn.

OK, here's what everyone is missing, and I think, an inherent problem with defensive statistics as currently constituted.

PNC Park has an odd configuration. The "North Side Notch" as it known affords LF's playing in PNC many more chances to get to balls that would otherwise be homeruns. Obviously, that allows a home team player more chances than opponents. Add to that, the odd defensive alignment the Pirates have been using all year (with Morgan shading way towards CF) and you get very skewed UZR ratings. Very skewed. Also, PNC Park very likely affected Jason Bay the other way, i.e., it exacerbated his shortcomings and the result was that he appeared to be a VERY bad LF instead of a marginally bad LF. Point is, a guy like Morgan with very good speed and the advantage of playing half his games in PNC Park ends up with a much higher UZR than LF's in Parks like Minute Maid or Fenway. At least, I think this is what the issue is.

Don't get me wrong, Morgan appears (by my own eyes) to be quite good. But, the system as the stats are figured simply can't account for these "oddball" factors I believe.

Stratman has it exactly right. Morgan is catching everything because he has been told to stand where the balls are falling. His stats are sort of analogous to a guy's batting average if he used an auxiliary batter's box 45 feet up the line.

The big question here is whether management created a statistical illusion on purpose.

Playing left field in Pittsburgh is playing centerfield anywhere else. It's a HUGE left field.

And the big steps forward of late by teams are situations where defense is emphasized. Tampa last season. Seattle looking competent this season. Pittsburgh helped itself by overhauling its defense and saw a guy like Zach Duke feel ok about throwing strikes again.

I think Adam Dunn got a raw deal in Cincy as Marty is a vindictive pr#ck. But the lad just can't get back to where he was five years ago after he let himself get sloppy fat in 2006. He is BETTER. But still a poor outfielder.

Morgan snarfs up everything like a starving beagle going after spilled food.

I buy this premise

PNC Park has an odd configuration. The "North Side Notch" as it known affords LF's playing in PNC many more chances to get to balls that would otherwise be homeruns. Obviously, that allows a home team player more chances than opponents. Add to that, the odd defensive alignment the Pirates have been using all year (with Morgan shading way towards CF) and you get very skewed UZR ratings. Very skewed. Also, PNC Park very likely affected Jason Bay the other way, i.e., it exacerbated his shortcomings and the result was that he appeared to be a VERY bad LF instead of a marginally bad LF. Point is, a guy like Morgan with very good speed and the advantage of playing half his games in PNC Park ends up with a much higher UZR than LF's in Parks like Minute Maid or Fenway. At least, I think this is what the issue is.

This is wrong:

1) UZR is park adjusted.

2) UZR is based upon the number of chances, such that an average player would (ignoring sampling issues) appear average regardless of the number of chances.

Morgan is catching everything because he has been told to stand where the balls are falling.

Where else would he stand? Do most other teams positions their outfielders in order to let more balls fall in for hits and make games more exciting?

Playing left field in Pittsburgh is playing centerfield anywhere else. It's a HUGE left field.

I'm not sure this is true. Morgan's speed was definitely an asset in LF in Pittsburgh. But his ability to judge balls in CF is a question to me. I am going to be interested in how things work out for Morgan in Washington.

I do not necessarily agree with Cameron in full, but I did say yesterday that I thought adding Morgan makes some sense for Washington--the Nats need to develop some pitching somehow some way and maybe adding OF D can help.

1) UZR is park adjusted.

Really? Then why is it understood to give screwy results for parks like Minute Maid (crawford boxes) and Fenway (green monster)?

Really? Then why is it understood to give screwy results for parks like Minute Maid (crawford boxes) and Fenway (green monster)?

Is it? Such as?

Y'all need to ask some NL Central left fielders about PNC. Soriano, Braun, etc have all commented.
Dunn would mock how he should try and sneak Ryan Freel out under his shirt to help cover left field in Pittsburgh.

As an example, Dial and others have found the UZR ratings for left field in Fenway to be way off. They are the experts here, so I leave it to them.

Y'all need to ask some NL Central left fielders about PNC. Soriano, Braun, etc have all commented. Dunn would mock how he should try and sneak Ryan Freel out under his shirt to help cover left field in Pittsburgh.

LF in PNC defintely is different and requires speed. Still not sure it equates to playing CF in any other park. Morgan definitely looked better in LF at PNC where his skill set seemed to fit perfectly. Like I said before..... I am interested in seeing how his skill set translates to CF in a different park. And I hope it works out for him.

As I recall, it's not that UZR has a problem with Fenway LF, it's that STATS does, because it was recording a balls high off the wall as potentially being catchable. I don't know if BIS has that problem.

I think the gaps between STATS-UZR and BIS-UZR are pretty discouraging, actually.

Dave raises an interesting idea in that their pitchers might develop better with better defense behind them. I have no idea if that's true, but it's plausible. Also doesn't seem like the hardest thing in the world to test.

Bill James made the same observation in the 1990 Baseball Book when talking about Chicago WS and Baltimore.

I completely agree with Cameron about Morgan and Dunn being pretty comparable in value in LF. (I'm pretty sure I made some blog comments to exactly the same effect when some Pirates fans wanted to sign Dunn as an outfielder.) The question, however, is whether you should trade anything of value for Morgan, since you can get something pretty close for nothing; Langerhans is a great example. The Nats were stupid to start three DHs in the outfield for all the reasons Cameron points out, but that doesn't mean they had to compound the error by trading away useful players to fix it.

The whole reason teams like Baltimore, Seattle, and even Pittsburgh have started emphasizing defense so much is that it is undervalued, so burning trade chips to acquire them defeats the purpose. That's why the Pirates made the trade -- they can easily find someone who's maybe half a win worse and costs them nothing. Heck, Jeff Salazar could probably do it.

It was on pp 49-50. James also mentioned Atlanta. Atlanta went on to have some success with their young hurlers after that.

Watching the Nationals at least forty times this year (painful enough) I really can't emphasize how bad their defense has been,and the effect it's had on the young pitchers. So I can't fault them for picking up Morgan. The idea that he is worth the same as Dunn is very wrong, though..the purge of the Bowden players bothers me not at all.

Playing left field in Pittsburgh is playing centerfield anywhere else. It's a HUGE left field.

But it's not in defensive statistics b/c of who you are compared with. Morgan looks great in LF partly b/c he is compared to guys like Dunn. In CF, the average ability is quite a bit higher.

I wonder if Gorky's Zygotic Hernandez could produce near-Nyjer-level performance sometime soon.

Trade McLouth because you think McCutchen is going to be just as good soon enough...
Trade Morgan because you think Hernandez is going to be just as good soon enough...

And you get prospects in the trades, too!

Of cours, GZH is slugging .278 in Altoona at the moment, but Rome wasn't built in a day.

Gorky's Zygotic Hernandez

Man oh man, do I hope this name sticks. I think the reference is too obscure though.

I think the gaps between STATS-UZR and BIS-UZR are pretty discouraging, actually.

I'm pinning my hopes on HITf/x to give better data.

It was on pp 49-50. James also mentioned Atlanta. Atlanta went on to have some success with their young hurlers after that.

Yes. I was going to say this. Everyone knew that Atlanta had a rather amazing array of young pitchers in the majors and the minors. In the winter of 1990 they made a series of what seemed to be extremely odd moves for a last-place team:

1. Signed 30-year-old Sid Bream, who was a mediocre-at-best hitter for a first baseman.
2. Signed Rafael Belliard, perhaps the weakest hitter of the past 25 years (OPS+ 46 vs 59 for Rey Ordonez).
3. Signed Terry Pendleton, a 30-year-old third baseman coming off a 601 OPS year.
4. Signed Otis Nixon, who was 32 years old and pretty close to being a 1980s version of Herbie Washington.

Basically every move was made to support the young pitchers. Every moves improved the defense, generally dramatically:

1. Bream put Justice in RF permanently, instead of at first base. Bream and Brian Hunter gave them good defense at 1B, and a marginally effective offensive platoon.
2. Nixon shifted Gant from CF to LF, and took at-bats away from Lonnie Smith, who was one of the team's best hitters in 1990(OPS+ 127) and was downright spectacular offensively in 1989.
3. Belliard reduced playing time for both Jeff Blauser and Jeff Treadway, both of whom could be described as "good bat for middle infielders," and finally got the dreadful Andres Thomas off the team.
4. Pendleton was a Gold Glover at 3B.

There is absolutely no doubt IMM that all of these moves would have been mocked if BTF had existed then.

Now, a lot of crazy things happened along the way, like Nixon and Pendleton hitting way better than anyone could have expected. But the pitcher staff took a dramatic step forward:

1. Tommy Glavine reduced his ERA from 4.28 to 2.55.
2. Steve Avery went from 5.64 with 121 hits allowed in 99 innings to 3.38 with 189 hits allowed in 210 innings.
3. John Smoltz was the most established of the young pitchers, and cut his ERA only slightly, 3.85 to 3.80, but he cut his unearned runs from 10 to 4.
4. The team's five main relievers (Pena, Stanton, Mercker, Berenguer, Freeman) all had very good years (ERAs ranging from 1.40 to 3.00).

Now can Washington replicate this? I don't know. They don't have Bobby Cox, for one. But Washington does have a fairly viable group of young hitters, and a reasonably talented young rotation. The team is being destroyed by it's historically awful bullpen and very bad defense.

I made a modest handful of questionable assumptions, ran some year-to-year correlations, and strapped together the basics of a UZR projection system. I come up with roughly a +13 in 150 games in CF, given my assumptions. Regression to the mean given the assumed chances involved should be just a touch less than 50%.

The whole reason teams like Baltimore, Seattle, and even Pittsburgh have started emphasizing defense so much is that it is undervalued,

I don't think BAL is emphasizing defense. Not when they move Huff from DH to 1B... or play Ty Wigginton at SS (granted he's only logged a few innings there)!

I have no idea about Morgan vs Dunn. I tend to mistrust UZR and other defensive stats, and I don't ever bother looking at them in small samples. But I have a different problem with the analysis:

Here’s the problem. Nyjer Morgan and Adam Dunn are nearly equals in value, and the polar reactions from the sabermetric crowd puts the blindspots that have been developed over the last 10-15 years on full display.

Statheads have come a LONG way since they were singing hosannas for the Grieve/Berroa/Becker/Stairs beer league softball style outfields in Oakland. Zone rating was a new thing then that wasn't easy to access. That was a different time and the groupthink has shifted significantly. In the last offseason, you had a whole group of leadfooted sluggers (Ibanez, Burrell and Dunn) generally meeting few suitors and there was no outcry on BTF. Back in the day when people were tearing their hair out just when they heard that Billy McMillon wasn't a good enough fielder to play in the majors.

The Morgan-Dunn equivalency mostly seems nuts because Dunn is a lock for 40 HR, 100 RBI, etc, whereas Morgan has never really accomplished anything. But you see similar suggestions ("I bet Endy Chavez produces just as many runs as Raul Ibanez") on this site all the time.

The difference between offense and defense is simple: there is a limit to the amount of outs for a defender to make (meaning, a cap on his defensive production) but there is absolutely no cap on offensive production.

This may have been discussed in the interim, so apologies if I'm rehashing here, but: This is not the difference between offense and defense. It's *a* difference, albeit a small and for practical purposes nearly meaningless one. Another difference, which I consider to be *the* difference (other than the fact that all involved agree that defensive metrics are not as refined/advanced as offensive metrics are, but I consider that a temporal problem, not structural), is that there is less stratification on the defensive side.

Now, this is partly due to that temporal problem, of course. However, it's a structural reality as well. Because of the much tighter true talent range of defensive ability amongst major leaguers, it's basically impossible to provide as much value over replacement or value over average on defense as one can on offense. Adam Everett's numbers even made mgl take a long breath and shake his head, and his UZR was what, right at 40, 41? Yet a 4 win over average offensive player isn't very hard to find at all.

We all know* it to be true, that it's pretty easy to find a guy who can field at 90% of the level of the average MLB player. It's not so easy to find a guy who can hit at 90% the level. A guy who can hit for (just plucking a metric) a 100 OPS+ and field 90% is much more valuable than a guy who can hit for 90 and field 100%. This has led to some arguments on the value side in the past, ones in which I've participated vociferously. All things equal, this means that a +20 on defense is more "valuable" than a +20 on offense, even while providing the same raw additions to wins. Roster and dollar scarcity dictate that one should pay more for the former than the latter if the market is perfect or somewhat close to cognizant of the factors.

To apply Stephen Jay Gould's evolutionary concept that I think was introduced in James' Politics of Glory, wherein it's posited that as organisms evolve and get 'better' the outliers wrt physical characteristics become more rare and less pronounced (you don't see 10 foot tall humans), you can look at the progress of fielding ability as it physically manifests in actual play and reverse the extrapolation. Gould used the lack of .400 hitters as a base to show that ballplayers are better now than in the past. We're fairly confident if not totally assured that fielding ability on the whole is much greater than it used to be, as such we can posit that the outliers should be more rare.

That rarity creates value where it might seem to not lie. The +10 on defense is equal to the +10 on offense, if the underlying metrics are correct. However, again, the +10 would be more valuable in a monetary sense. *That's* a, if not the, difference.

To apply Stephen Jay Gould's evolutionary concept that I think was introduced in James' Politics of Glory

I'd bet that this concept was not introduced in any of James' works. Referenced maybe, but not introduced. (Dumb nitpick, I know. Shoot me.)

My real question, is why is there more variance in hitting than in defense?

My real question, is why is there more variance in hitting than in defense?


Remember, the way positions are assigned exaggerates the difference. Players are essentially sorted to position on the basis of defensive quality, so the differences in runs prevented are going to look small.

My real question, is why is there more variance in hitting than in defense?

Because even you could catch a major league hitter's fly ball for an out SOME of the time. You could probably even snag some ground balls.

But you'd be exceedingly lucky to do more than strike out, even against the worst MLB pitcher.

The number of opportunities for a player on defense is so much smaller, especially when you weed out the plays that any MLB player can make at least 80% of the time and focus on the relatively small number of plays where players can meaningfully differ in defensive abilities.

Bill James made the same observation in the 1990 Baseball Book when talking about Chicago WS and Baltimore.

That is where I got it, and I think it makes a lot of sense. Also, the 1991 Braves focused on D.

I'd bet that this concept was not introduced in any of James' works. Referenced maybe, but not introduced. (Dumb nitpick, I know. Shoot me.)

You're of course correct in that I meant introduced as in "referenced, and likely the first reference that the vast majority of readers had seen" in PoG. Shooting you remains an option.

My real question, is why is there more variance in hitting than in defense?

I can think of a few reasons, off the top of my head and therefore subject to wild swings in their ability to satisfactorily explain the phenomenon and their ability to make sense (not to mention my belief in them):

1) The pitch vs. the batter is a human controlled event on both sides. A pitcher throws low and outside to a hitter who doesn't hit that pitch well. This being an MLB pitcher, he's much more often than not able to get that pitch low and outside. A fielder who is marginally worse at handling line drives to his left above waist-level doesn't face hitters who are able to consistently exploit this deficiency in the way the pitcher can.

2) Something physiological is 'easier' about muscle-memory and the like affecting fielding vs. hitting, likely related back to the same theory as #1. Essentially, 'practice makes perfect' works better for fielding than hitting.

3) Equipment advances. Bats have stayed much the same lo these many years, orders of magnitude moreso than gloves. The advance in glove technology may exactly cover the difference in fielding skill that we would see if all fielders went back to 1930s gloves.

4) Self-selection. This is obvious, but it's quite possible that marginal fielders are selected out of the population at a higher rate early on in the process, at a greater level than hitters. A bobble or bad throw from a JV shortstop is much easier to spot than the 1 at-bat a week separating a .300 hitter from a .250 hitter. It's easier to pick out the bad fielder immediately than it is the equally poor hitter.

5) Also related, swing mechanics are much more personalized and therefore the study/evaluation of them is much more complex. This is somewhat circular, but basically #4 and #5 would combine to weed out the marginalia of fielders before hitters.

A guy who can hit for (just plucking a metric) a 100 OPS+ and field 90% is much more valuable than a guy who can hit for 90 and field 100%.

Before anyone jumps on this, I was very imprecise with my word choice here, especially considering what immediately followed. The first guy is much harder to find than the second guy, is what I meant. The original sentence is potentially conflicting with the rest of my paragraph.

I just want to know why Dave Cameron is defending himself here but not on his own blog.

If Morgan is a plus CF for the next three years it's well more than likely that he'll be more valuable over his career than Milledge ever will be.

So how much more valuable will Dunn become once he's moved to first? Is there any way Morgan's value could be increased by the same amount - that is to say, how do you move him rightward on the offensive spectrum?

how do you move him rightward on the offensive spectrum?

Hit him cleanup, so he gets more ribbies.

So how much more valuable will Dunn become once he's moved to first? Is there any way Morgan's value could be increased by the same amount - that is to say, how do you move him rightward on the offensive spectrum?

I was kinda wondering the same thing... Say Morgan is a +4 Win player as a LF and Dunn is also a +4 Win player as a LF, but a +6 Win player at 1B (to throw out random numbers). At what point is it not an issue with Adam Dunn but an issue with management?

Wasn't Adam Dunn tried at first base by the Reds and the results not exactly good because his footwork was a mess?

I was kinda wondering the same thing... Say Morgan is a +4 Win player as a LF and Dunn is also a +4 Win player as a LF, but a +6 Win player at 1B (to throw out random numbers). At what point is it not an issue with Adam Dunn but an issue with management?

Well, it's always something of both, of course.

1) It's always an issue with Adam Dunn because in the end, it is about his inability to be exactly what the Nats want in their most perfect of worlds. As it is with every player. His skillset and his application of it are what defines his value, no matter what context or framework, so he is always the issue.

2) It's never an issue with Adam Dunn because in the end, it is about his skillset and his application of it. Absent charges of jaking it, how can it be Adam Dunn's fault that he isn't some superhuman demigod? He is what he is, if he is a professional and gets what he can out of his talents, you can't blame him for being subpar in left field anymore than you can blame him for not writing like Twain or not being the fifth member of the Ninja Turtles.

Regarding the limits of performance mentioned earlier: you are right there is a limit to how many runs a player can save. But the amount of extra runs a bad fielder can allow is much greater than this. On offense, there is no limit to how many runs a team can score, but there is a limit to what an individual can add to a team, which in the case of the perfect batter the runs he'd add by being walked all the time. I don't think any of that is relevant to what a +20 hitter or fielder is worth.

I was kinda wondering the same thing... Say Morgan is a +4 Win player as a LF and Dunn is also a +4 Win player as a LF, but a +6 Win player at 1B (to throw out random numbers). At what point is it not an issue with Adam Dunn but an issue with management?

For that to be the case, you'd have to expect that Dunn makes the transition from LF to 1B substantially better than the majority of players would. Position adjustments are set up so that, on average, a player who moves from one position to the other is equally valuable.

That "on average" is the sticking point of course, but I haven't seen anything to suggest that Dunn would be one of the players who is above the curve - let alone 2 wins above the curve.

Wait, wouldn't the perfect batter homer every time?

Can he hit the intentional pass 15 feet outside while keeping his feet in the batter's box? Then yes, he'd homer instead of taking the IBB.

Reed Richards, yuh huh

I think this is a win/win move, and have since the moment I saw it. I really should have posted in one of the earlier threads on it but the time wasn't available. The baseball analysis community which feels the need to decry winners and losers to each transaction is a faulty analytical frame work, which I have been saying for years. Hopefully, one of these days more people which catch on.

Now away from META, the reason I like this trade from the Pirates perspective is Hanrahan, who I think is very likely to turn into a quality back of the bullpen option. The type of guy that teams really value. If you think teams don't value that, the Mets could have dealt Eddie Parnell earlier in the year for legit big league talent, because he was a big hard throwing right hander who had 15 great innings.

The reason I think the internet baseball analysis community is going overboard on the positives of this trade for the Pirates is because they still say significant potential in Milledge, a player I have panned for three years on this site because I didn't think he took baseball as serious as he needed to develop his talents into skills. The Washington experience has added additional evidence to help confirm that belief. I very much liked the trade that sent him to the Nationals, as I felt that was a perfect destination for him to turn the corner. Yet, he continues to flounder. I would be surprised if he turned into a player.

All things equal, this means that a +20 on defense is more "valuable" than a +20 on offense, even while providing the same raw additions to wins

YES. JEFF K. = SUPERSTAR!!!!!!!!!!!! I've been arguing this since my attempt to boost Mazeroski's Hall of Merit votes a couple of years ago.

Is exceptional defence + average-for-position hitting underrated? It seems it to me, but I may only be remembering those who underrate it.

I'm skeptical that Langerhans could have done the same job as Morgan. Looking at his UZR section on Fangraphs, he has never played especially well in center, and his defensive reputation is built off excellent play in the corners for the 2005-2006 Braves. He's essentially average for 2007-2008. He's a big guy (6'3, 229) without blazing speed, so I think it's reasonable to think he might not be as good as he was 3-4 years ago. That he was playing a lot of 1B in the minors doesn't help the assessment.

Preseason projections:

Morgan: -22 offense. +15 LF, +10 CF

Milledge: +0 offense, -1 LF, -10 CF

If we say Morgan has improved his offensive projection to about -15 and Milledge unchanged (generous assumption), then he's -5 overall in center (before position adj), and Milledge is -10.

But in left field they are closer, Milledge is -1 and Morgan is 0. Because a centerfielder gets more chances, you can easily find a situation where player A is more valuable than B in a corner, but B is more valuable than A if used in center. Nat's needed a CF, and Milledge is their 4th best corner OF. Pittsburgh is set in center for probably another 5 years (before the need to trade McCutchen for prospects) and could use a better bat in the spot. Both teams should be happy.

[90] Interesting comments. Definitely agree on Hanrahan -- who most Nats fans were ready to torture.

One thing that often goes missing from these sorts of evaluations, too, is any sense of how it addresses a team's strength or weaknesses. Most evaluations are simply on the basis of talent versus talent.

But trades don't happen in a vacuum. Teams have strengths and weaknesses and certain players may have less value to one team than they do to another.

Sometimes there's accounting for that, but it's hard to quantify that in a way you can by figuring out a player's WAR or whatever.

In this case, the Nats have 0 credible MLB-ready CFers. And they have 102 MLB-ready corner outfielders. Even if the value-for-value aspect of the trade is a net loss for the Nats (and I'm not saying it is), it's likely to improve the Nats because Milledge is a player they weren't able to play given their depth at the corner positions.

Whether Morgan is a +10 CFer or a +5 is part of that debate too. Those are all comparisons to average performance. Well, the Nats aren't getting average out of CF. They're getting -20.

How do the pieces of a trade fit together? How do trades address organizational strengths or weaknesses? Those are the things we need to look at, besides just the individual player's numbers up on fangraphs.

Player A: 278/360/405
God: 247/297/388

Player A: 278/360/405
God: 247/297/388

Gotta trade God for player A, no?

I'm skeptical that Langerhans could have done the same job as Morgan. Looking at his UZR section on Fangraphs, he has never played especially well in center, and his defensive reputation is built off excellent play in the corners for the 2005-2006 Braves. He's essentially average for 2007-2008. He's a big guy (6'3, 229) without blazing speed, so I think it's reasonable to think he might not be as good as he was 3-4 years ago. That he was playing a lot of 1B in the minors doesn't help the assessment.

Good point on the 1B thing. But I think that that had as much to do with the team's depth in the outfield. They viewed Langerhans as a bit of a spare part -- a stopgap kind of guy. As such, he played where the "prospects" weren't, so where he played might not be the best consideration of his defensive value.

That being said... the comments about LF/CF are interesting. I noticed when looking at the numbers when he was playing regularly that he seems to benefit from playing with an aggressive shift.

At least 'til this last year or two, the Nats (thanks to Pat Corrales, who did the same thing with the Braves) shifted their outfield defense very aggressively.

To my eye (and the way I think some of the D metrics work -- i know this was the case with some of the Nats RZR, for example) they were getting to balls nobody else got to not necessarily because they had range, but because they were 10 feet off the line to begin with.

I'm not sure how UZR accounts for dramatic shifts, or even if it can. But if a team has a coach (as the Nats and Braves did when Langerhans was there) who sets their positioning, their "range" might not translate to a team that doesn't similarly position their players.

At the very least, you figure an omniscient god would have a better batting eye.

At the very least, you figure an omniscient god would have a better batting eye.

But God is all merciful so He may make outs out of pity for the pitchers.

God would be all about the sacrifice bunt. He'd be hated here.

Yeah, but you know how tie goes to the runner? Think about God.

Besides, he'd have the smallest strike zone imaginable.