Ugh. I don't really disagree with the premise, but wouldn't it have been a good idea, to pick a player, where we have some confidence that his fielding #s aren't complete bunk because of the green monster?

It won't fly if you just say "-15 runs", but if you go find specific plays where Manny:

- should have made the play
- didn't make the play
- the plays led to runs
- the runs add up to -15.

then people will believe it. It's time consuming, but would be effective, I guarantee.

Or, if you look at all the plays and can't find 15 runs that Manny cost his team, maybe the metrics aren't right.

Manny's numbers, unadjusted for the monster, would be more like -30 to -40.

At -15, Eric is saying he's an equivalent fielder to Dunn, Burrell, Ibanez, or Carlos Lee. Is that conclusion really controversial?

Manny's numbers, unadjusted for the monster, would be more like -30 to -40.

At -15, Eric is saying he's an equivalent fielder to Dunn, Burrell, Ibanez, or Carlos Lee. Is that conclusion really controversial?

I wonder about this based on Manny's play in LA and his first year in Boston. This isn't a good thing to say, but I think after Ortiz became "The Man", Manny got slack in the OF. His defense really slid. He may be -15, but I could see him being -8. One thing that tipped me to this was the comment that "Manny was just about as bad on the road." BEFORE taking the Monster into account.

So I think Manny was not hurting himself out there, and if he signs with a team that says "Manny is THE MAN", he'll be a monster at the plate and fine (not average) but fine in the OF.

I agree with what Dial's brief assessment here. And unlike me, he actually knows what he's talking about when it comes to defensive metrics.

However, I do think that limits the places that are good for Manny to sign, *and* I think Manny will show a preference to go to a team where he isn't likely to be second fiddle. That could be the Dodgers. I htink he's less likely to sign with the Mets or Yankees because of that. And I think if he does sign because hte money is too awesome there would be a risk in a year or two that he'd be "Manny being Manny". I really think that stuff was to get atttention he thinks he deserves (and he does - he's the best player on his team, with very few exceptions).

Now, if I were to go on air at ESPN or write a column at an extreme mainstream site discussing how Crawford in 2009 is almost just as valuable, all told, as Ramirez, can you imagine the type of criticism I would receive.

Even with all of the defensive metrics available today, this is an area that still isn't concrete from an evaluative standpoint, which makes it tough to justify its importance when statheads like us are under intense scrutiny to mathematically and logically prove our points simply to gain an oodle of acceptance in the mainstream.

Perhaps if a defensive metric comes out that leverages the importance level of the situation we could begin to show this, or one that, with very educated estimates, shows what hit would have been recorded had the ball not been fielded (as in, factoring in the run values of what would have occurred had the play not been fielded), but this is still a tough analytical facet to sell.

This doesn't show up in the form of errors or fielding percentage, though; even though their are innate flaws in new age defensive metrics and our perceptions of their usage, we can all agree that fielding percentage is awful.

Seidman makes some interesting points in this article, but I always find his writing painful to read. A little editing would go a long way.

Hey guys, from the discussion we had over at Fangraphs, it seems like another big key here is that defense varies much more than offense, and perhaps Crawford wasn't the best comparison due to this, but there are a plethora of players we could have used to show why he isn't as valuable as many may think due to his poor defense. Then again, in terms of money, which this wasn't about, it could be argued that he puts butts in the seats, and would be worth a deal overvaluing his offense because of this. I was really just trying to figure out the reasons behind why we undervalue defense. So far, from the post and discussion, we have:

1) Our perceptions get warped, IE - Beltran gliding to a flyball while Victorino scurries to the same ball. It makes Shane look better.
2) We get spoiled from watching great fielders and evaluate things in terms of whether or not the play was made, not necessarily how incredible the range was to even have a shot at the ball.
3) Fantasy Baseball is incredibly prominent and doesn't include defense other than errors on occasion.
4) Defense is contingent upon pitching, the two are interlocked much more than offense.

Anything else you guys can think of? The first step to eradicating a problem is figuring out what causes the problem.

DKDC, thanks for pointing that out. I wasn't quite "with it" last night, and managed to go back and edit it just now. I can't believe I wrote some of that stuff in the ways the sentences were written!

Or, if you look at all the plays and can't find 15 runs that Manny cost his team, maybe the metrics aren't right.

That's not necessarily true. Offensive statistics are the same way. The total value of all of Manny's hits might not equal +30 runs every season. Statistics don't work that way.

I've been saying for a couple years now that Crawford has actually been more valuable than Ramirez in recent seasons, though that wasn't the case in 2008 with Crawford's struggles at the plate. Normally defense and baserunning won't erase as large of an offensive gap as Ramirez's high level of production with the bat creates but you're talking about the difference between the best defender at the position and one of the worst, as well as a guy who creates a lot of value with stolen bases and speed in general vs. a slow guy with poor baserunning skills.

To clarify, I'm aware that the baserunning numbers as far as steals go are already part of the offensive numbers, my point was that the gap would be larger than it already is if you based it off of bat alone.

@ Eric: I'd like to add to the list. How about:

5) We like to believe that awesomely talented athletes are awesomely talented all-around. This might help to explain why so many Gold Gloves go to awesome hitters who are fair to middling with the glove. Jeter can hit for average, has some good pop, can run the bases, can steal, has great awareness, and looks very fit and athletic. Therefore he must also be agile, versatile, and aware on defense.

6) The farts & fireworks principle: My friend's dad once told me that the two things that will always make you smile are farts and fireworks, regardless of how jaded you may be. When Jim Edmonds dives for a ball in the gap, when Jeter leaps into the air and fires the ball to first, when Griffey Jr literally glides right up the kingdome wall to snag a homerun, we all smile. Those momentary explosions are far more powerful and evocative than, say, Gary Pettis or Curtis Granderson or Adam Everett simply getting to everything and recording tons of outs.

#6 may also be #1...

it seems like another big key here is that defense varies much more than offense, and perhaps Crawford wasn't the best comparison due to this

This is a good point - defense does vary more - but it is also much more difficult to measure, and estimates of defensive valuelag far behind estimation of offensive value in terms of their precision. All sorts of reasons for this - interaction between fielders, park effects, pitching, lack of resolution in data, lack of resolution in estimation techniques, etc. You can say a guy was "-15" in a given season, but how much uncertainty does that estimate come with? IMO, too much to use it as is, especially for a single season.

This is not meant as a slight on those who have beenn working on freely available advanced defensive metrics - these have come a long way in just a few years, and I'm not saying they aren't useful - I'm just saying that we shouldn't in any way be equating them with estimates of offensive values, because the uncertainty underlying them is likely far too great.

It won't fly if you just say "-15 runs", but if you go find specific plays where Manny:

- should have made the play
- didn't make the play
- the plays led to runs
- the runs add up to -15.

then people will believe it. It's time consuming, but would be effective, I guarantee.

Or, if you look at all the plays and can't find 15 runs that Manny cost his team, maybe the metrics aren't right.

In one of his books, Bill James wrote about working on George Bell's arb cases a couple of times and putting together an exhibit that showed that Bell's errors--he made like 10--actually didn't really cost the Jays any games.

Now, I know "runs" are not games, but the quoted post made me think of some thing I have been wondering about, which is "leveraged defense." IOW, if Ramiriez does not get to a ball with two guys on and it rolls to the wall when the Dodgers are up 7-1 and they win 8-3, that is different than if he doesn't get to one when they are tied 3-3 and they lose 5-4. This is of course why defensive replacements are used in close games, but I wonder in what way this kind of thing is taken into account in defensive analysis.

According to Dewan, with the "Monster Adjustment" Ramirez's +/- was -24 last year. Matthew Carruth, in writing about the Rays in THT, said each +/- = about .8 runs. So Ramirez, by those metrics,is costing his team about 19 runs, or about two games a year, by his comical stumbling around in LF.

But I wonder, if you analyzed it using the idea in #2 and approached it as James approached Bell's errors, if it is really that much, or more,or less, in terms of the W-L column.

In one of his books, Bill James wrote about working on George Bell's arb cases a couple of times and putting together an exhibit that showed that Bell's errors--he made like 10--actually didn't really cost the Jays any games.

That story was in the New Bill James Historical Baseball Abstract (2003).

On not valuing defense ...

An interesting thing is that a few years ago, (most) statheads placed little value in defense and were criticized by the mainstream for it (all the "beer-league softball" comments). Now (more) statheads have come closer to the mainstream's point of view ... and, well, I don't see them necessarily meeting much resistance from the mainstream on this. The mainstream might generally use different (poorer) measures like FP and errors, but Dewan's Fielding Bible was quickly picked up by the mainstream (the power of a good title!) and, with the high profile exception of Jeter, given a lot of credence.

So I don't know that you'll find all that many folks in the mainstream who won't buy that average offense and great defense are as valuable as great offense and poor defense. And in the case of Manny, whom many of the media love to jump all over, I think you'd find a lot of takers for the notion that Crawford is as valuable overall. This would be especially true if Crawford hadn't hit like crap this past year and Manny hadn't "singlehandedly" saved the Dodgers. (i.e. part of the "problem" at the moment is that Manny put up a 219 OPS+ in LA and I don't care if he was Dick Stuart in CF with the glove, he couldn't give back that many runs)

I am not sure defense is really "undervalued" in the mainstream ... it would still seem to be improperly measured in the mainstream but, like the strikeout and "clutch" hitting, it's still widely cited as counterbalancing (not just reducing) the offensive contributions of the Dunns and Burrells of the world. But there are obvious reasons it doesn't get that much attention unless you're Ozzie Smith. A hit is clearly a hit and (more to the point) a RUN scored and a RUN batted in are clearly RUNS, a GAME-winning hit is clearly a GAME-winning hit and there are no defensive counting stats of importance. The counting stat nature of errors is probably why they're so attractive but, for most positions, we're talking about differences like 10 vs. 15 errors not 127 RBI.

Give defensive stats that provide counting measures of defense (plays, runs) another decade or so and I think you'll see them more widely acknowledged and given credence in the mainstream. And it's fairly clear that many front offices have started doing something like this and that means that eventually it will disseminate into the media.

I think the main problems are not having a single agreed upon standard and a lack of access to the numbers. If Dewan's plus/minus were expressed as a run value and listed at ESPN and Baseball Reference, the problem would take care of itself.

I'm also not sure that using Derek Jeter's shoddy defense as a main vehicle to promote defensive statistics has proved a very good idea. People are less likely to listen when they're being told that they're wrong and possibly stupid.

i have always thought that fantasy baseball not including defefnse as valuable has made some difference.

also, the belief that moneyball/stats/sabr show that what matters is only walks, K, HR and basically OFFENSE, has made a lot of minds up.

that and no way main publishing of runs saved vs runs cost ONLY on defense

i couldn't never convince most people that adam everett was a positive on the team even if he did have a .240 BA because his defense saved so many runs. but lots of people really REALLY do not believe that bad defense costs any runs (see clank lee) and anyhow, it is ok since he doesn't K much...

- and agree with tim m about the trouble with using saint jeter as THE example of crappy fielding. that turned lots of people OFF just like trying to persuade them that a guy who hits .240 while saving more runs than ozzie was actually valuable - after all, he only hit .240 so how could he be valuable?

This is not meant as a slight on those who have beenn working on freely available advanced defensive metrics - these have come a long way in just a few years, and I'm not saying they aren't useful - I'm just saying that we shouldn't in any way be equating them with estimates of offensive values, because the uncertainty underlying them is likely far too great.

I entirely agree. And this is the flaw in Eric's run projection comparison between Manny and Crawford: it's (at this point in history) a fallacy to consider the offensive and defensive elements in the equation as equally reliable.

I've always thought of the defensives stats as pointers rather than absolute measures. E,g, if a bunch of them are all "-ve" then it's a good indicator that, hey, this guy's not so great in the field.

But there are obvious reasons it doesn't get that much attention unless you're Ozzie Smith.

At his peak Ozzie might have been a 5+ Win above replacement player - one of the very best in the game. Take average offense or so from a SS and add 30 runs with the glove. And I think Ozzie was indeed one of the highest paid players in baseball. I don't know if an Ozzie type, playing today would make 25 million though.

Manny should have no trouble covering left field...if the Dodgers move back to the Coliseum...

4) Defense is contingent upon pitching, the two are interlocked much more than offense.

I disagree with this. Offense is just as interlocked with pitching, and probably moreso.

I think that one reason that defense is undervalued is that fielding is not the skill that makes someone a professional baseball player. There are (relatively) a whole lot more people who could, with practice, field as well as the average major leaguer than could hit as well as the average major leaguer.

4) Defense is contingent upon pitching, the two are interlocked much more than offense.

I disagree with this. Offense is just as interlocked with pitching, and probably moreso.

In one sense yes. But over the course of a season, hitters hit against a wide mix of different pitchers, whereas defenders play behind a much smaller number of different pitchers.

So it seems to me that, as a first approximation, it is reasonable to treat hitters on different teams as if they've faced roughly the same level of opposing pitching over the course of a season (since they have faced a reasonably large sample of the league's pitching). On the other hand, it doesn't seem so reasonable to make a similar approximation for defense. So defense and pitching are more interlocked in that sense.

Apologies if I've missed the point, or if I'm saying something obvious.

But over the course of a season, hitters hit against a wide mix of different pitchers, whereas defenders play behind a much smaller number of different pitchers.

They face the same pitchers, but they do not face the same pitches. What Jeff Francouer gets thrown isn't the same set, even against the same pitchers as anyone else (well, except Andruw Jones). So when we talk about what a batter hits, we are completely assuming, since they face the same set of pitchers, they get similar pitches to hit. They do not.

When someone can produce evidence that fielders behind particular pitchers see a decided advantage, I can buy this as a signifcant effect. As I have studied (albeit through data prxy), I haven't found an apparent advantage. Yes, they are linked, but I believe studies show that once the pitcher releases the ball, whether or not it becomes a hit is more reliant on the batter (and fielders) rather than the pitcher.

I'm also not sure that using Derek Jeter's shoddy defense as a main vehicle to promote defensive statistics has proved a very good idea.

I think it has been a terrible idea. Sticking a controversial statement out there among the general baseball-watching population and declaring that you're right and people should listen to you isn't going to work very well. Defensive stats would be more easily accepted if their backers said something like "These numbers show that Adrian Beltre has lived up to his contract for the Mariners despite the drop in offense" or "Carlos Lee is providing a strong bat for the Astros but I can show that his defense is making him seriously overpaid". It's more effective to start out with something that most people can readily accept (and may be thinking themselves already) and go from there.

Sure if your goal was acceptance of the thesis rather than book sales.

Sure if your goal was acceptance of the thesis rather than book sales.

But does "Jeter is a terrible fielder" really sell books? I suppose there are a certain number of people who buy the books they disagree with hugely--GOPers who invest in Al Franken or Liberals who buy Ann Coulter and the like--but generally if I see book advocating a point I see as purely outlandish, something like one of the 9/11 Truther's which is how I imagine most people view calling Jeter a bad defensive SS, I wouldn't buy it.

But does "Jeter is a terrible fielder" really sell books?

I don't think the unconverted are really the target audience for these books. While I'm sure BPro and others would love to gain a greater foothold with the masses, their audience is surely guys like us.

1) Our perceptions get warped, IE - Beltran gliding to a flyball while Victorino scurries to the same ball. It makes Shane look better.

Really? It makes Shane look worse, if you ask me.

But over the course of a season, hitters hit against a wide mix of different pitchers, whereas defenders play behind a much smaller number of different pitchers.

They face the same pitchers, but they do not face the same pitches. What Jeff Francouer gets thrown isn't the same set, even against the same pitchers as anyone else (well, except Andruw Jones). So when we talk about what a batter hits, we are completely assuming, since they face the same set of pitchers, they get similar pitches to hit. They do not.

But the pitch set that Jeff Francouer sees against a particular pitcher depends on Jeff Francoeur's hitting skills. For example, if Francouer flails at a certain type of pitch, it's his fault if he sees more of it. So I think one can still say that Francouer is getting (essentially) the same overall opportunities as other hitters, even if manifested in different pitch sets.

On the other hand, I don't see how a defender's skills really influence what sort of fielding opportunities he gets.

When someone can produce evidence that fielders behind particular pitchers see a decided advantage, I can buy this as a signifcant effect. As I have studied (albeit through data prxy), I haven't found an apparent advantage.

Fair enough. (Although the effect wouldn't necessarily be that [all fielders behind particular pitchers see a decided advantage], merely that [fielding behind a given pitcher changes the relative value of different defensive skills]. So some fielders might look better behind particular pitchers, while their teammates might look worse behind the same pitcher.)

This challenges my intuition, though. For example, if I'm a 2Bman who is above average at fielding grounders and below average at going back on shallow fly balls, wouldn't I have a better +/- playing behind a groundball pitcher than behind a flyball pitcher (all other things being equal)?

Yes, they are linked, but I believe studies show that once the pitcher releases the ball, whether or not it becomes a hit is more reliant on the batter (and fielders) rather than the pitcher.

But [whether/how much a pitcher affects the frequency that balls in play become hits] is different than [whether/how much a pitcher affects which defensive skills are required behind him].

Again, consider a 2Bman who is above average at fielding grounders and below average at going back on shallow fly balls. His +/- would be helped (wouldn't it?) when playing behind a groundball pitcher, not because the pitcher is controlling the BABIP, but because the 2B's skills complement the pitcher's skills.

Doesn't that make a +/- rating more contextual than hitting stats?

Or am I missing something?

(Sorry if I'm belaboring this. And thanks to Mr. Dial for responding to my earlier comment.)

Really? It makes Shane look worse, if you ask me.

I imagine that's because you're impressed by innate ability, not by effort.