The question for me is at what point do your returns diminish with defensive adds? We were already first in DER, so how much of an improvement can be had? How much more can be squeezed out of the defensive end of the equation?
It becomes an obsession instead of an undervalued commodity when your returns fall off considerably.
The Rangers used to try to win everything with offense, but at some point you need to be able to pitch against the other guys. We already play D - what we don't do is score. The Yankees won a WS when they added defense to their incredible offensive machine. Without some offense, we're not winning anything, even with all the gains we can possible harvest from defense.
I don't know where the tipping point is on the tradeoff from scoring to defense, and the adage is that a run saved is equivalent to a run scored, but at some point the defense is so tight that almost all the runs that can be saved ARE being saved, and there's nothing left to harvest from a defensive perspective.
And at that point we'll need to add runs by actually SCORING them - whether or not those scored runs cost us more money to obtain. Refusing to add offense and instead going for the cheaper add of more glove savvy then does become an obsession instead of a value choice.
~G
Doc sez: The M's prejudice in favor of slick defense gets awfully annoying to Dr. D at times ...
Interesting to me that even with the results of 2009, very bright people continue to dismiss the value of defense as if it is some trivial component of the mix, of (obviously) lesser importance. Here are some facts about 2008 to 2009.
- In 2008 - Texas and Seattle finshed 14th and 13th in DER in the AL.
- In 2008 - Texas and Seattle finished 14th and 11th in runs allowed in the AL.
- In 2008 - Texas and Seattle finished 13th and 11th in FIP in the AL.
- In 2009 - Seattle and Texas finished 1st and 3rd in DER in the AL.
- In 2009 - Seattle and Texas finished 1st and 5th in runs allowed in the AL.
- In 2009 - Texas and Seattle finished 11th and 9th in FIP in the AL.
Note: In 2009, Seattle beat league avg. DER by 24 points - Texas beat it by 9.
Texas added 8 wins ... Seattle 24. Texas dropped from #1 in run scoring (in 2008) to almost exactly league average in 2009, (ranked 7th), while winning 8 more games. In NEITHER case, did the pitching improve appreciably. Seattle dropped from 13th to last (4.14 to to 3.95 runs per game -- and won TWENTY-FOUR (24!!!) additional games.
And completely lost in the shuffle - in 2008, the YANKEES were 12th in DER, missed the playoffs for the first time in ages -- but in 2009 ... the YANKEES (the team the won the WS), were #2 in DER. And they were also #2 in runs allowed.
Tampa, meanwhile went from #1 (.708 DER) to #4 (.695 DER), and lost 13 wins, (and failed to reach the playoffs).
The Seattle *DEFENSE* was so good in 2009, that the club climbed over EIGHT (8) teams with better pitching, to lead the league in ERA and run prevention - and they did so by a WIDE margin, (0.30 runs).
This is classic Moneyball. Defense *IS* undervalued. The 2009 club cut payroll while improving defense drastically, and added 24 wins. Tampa's defense slipped just slightly, and they misplaced 13 wins, despite the fact their offense jumped from 774 to 803 runs scored. Buying HRs is expensive. Buying OBP is expensive. Buying defense is cheap.
But, beyond that reality is this --- that it is likely (IMO), that a significant chunk of defensive results are produced solely by DESIRE to be good defensively. The focus is on identifying the best athletes. Who is fast? Who is smooth? But, when you're playing defense for 13,000 - 15,000 PITCHES per season, to consistently get that good jump, you HAVE to be paying attention for all 15,000 pitches -- to get that edge those 800 or 600 or 400 times it matters for your position.
Well, I believe it is IMPOSSIBLE to instill that desire to be vigilant defensively, if you make moves which 'clearly' place more value on offense. To create a foundation of defensive IMPORTANCE, you must make defensive vigilance a priority for everyone - regardless of their inate defensive ability.
AFTER you have instilled that paradigm, then if you bring in a weak defensive player, HE becomes the one needing to work to live up to the standards of the org. But, you've gotta be serious about it -- you've got to be willing to sacrifice offense (in a given situation), to prevent defensively laziness from spreading. Because, it WILL.
I believe Tampa's drop in DER from '08 to '09 is likely attributable directly to "relaxing" on defense, just a bit.
Yes, the Ms need more offense. And I'm one of Carp's biggest fans. But, sacrificing defense for offense is a dangerous proposition, because of the potential for the OTHER 8 defenders who might not take their own defensive vigilance as seriously. While it "can" be done - it is no easy task. With only a single season of defensive excellence behind them, I do not trust that one could blithely sacrifice defense in a significant way, and hope to retain the excellence of 2009.
The truth is that defense is responsible for 20 out of 27 outs per game. Yet, the pitching-first crowd will blithely state that most of "those" outs are "easy". Yet, EVERY strikeout is earned. I don't buy it. Any pitcher with a passable slider gets "easy" strikeouts against Beltre every bit as easy as the slow roller to Kotchman.
Defense is the easiest portion of the game to "slack off" w/o it being immediately obvious. And this is likely why league DER tends to drop slightly in the 2nd half. And it's also why winning teams seem to have their DER trend upward, while losing teams see their DER trend downwards.
It's kinda sad that the team can gain 24 wins in a single season, attributable almost entirely to improvement in defense, and yet the emphasis can still be viewed as obsession.
Comments
I think that statistical analysis has demonstrated that offense is primarily a series of independent contributes, but the issue is still up in the air for defense. If each defensive contribution were primarily independent, then there shouldn't be any diminishing returns. If you improve your shortstop, how good or bad the third baseman is has no impact on the upgrade from Yuni to Jack Wilson. If, however, the zone of influence of the third baseman and the shortstop overlap significantly, then the fear of diminishing returns certainly applies.
I'm agnostic on the subject.
That this team needs to score more runs, a LOT more runs. And I think that the 2010 offense WILL score more runs. But what I think what Sandy is saying is ALSO very true, that EVERY decision you make HAS to take defense into consideration, otherwise the tendency will be for defense to slide. Even if the intention is for it NOT to slide, by simply not emphasizing it, even the same set of players will tend to "let up" on defense over time, Tampa Bay from 2008 to 2009 for example. Largely the same players, yet they got more concerned about their offense than their defense, and viola. So if sacrificing a few runs on offense to take Chone at 3rd instead of, say Lowell, or Kotchman at 1st instead of Adam Dunn, then it is worth it to keep the entire teams mindset FOCUSED on defense. And then when we DO bring in AGone, or Prince in July they will know by the clubhouse that they must concentrate on defense every single pitch.
It isn't just about preventing more and more runs it is about keeping the run prevention we already have. And for that reason as long as JackyZ is at the Helm you can cross off all the Dunn's and Dye's in the world (unless they come here to DH).
Great article, Sandy, cPoints to ya.
The way I think of where to draw the line at diminishing returns for any skill is to think about pythag logic. Because the game, for lack of a more convenient description, plays pythagoreanly in marginal run scoring.
If you allow 500 runs (and have league average - 765 RS - offense) you'll win 113 games.
If you allow 100 MORE runs with the same offense, you'll win 100 games. That's 13 fewer wins and 100 fewer runs saved (you'd predict TEN fewer wins...so there's a drop-off in expected returns, but not a MASSIVE drop-off)
It's not impossible to allow 500 runs. It just hasn't been done recently because no team has ever tried to build a massive juggernaut defense
I wouldn't project that the Mariners have gone THAT far...but I don't think we're at the point where the returns for our efforts have become useless.
Valid question, G. But, what evidence is there that there IS a diminishing returns in concentrating on defense? I cannot find any. If anything, the data on DER may indicate an accelleration in winning associated with run prevention. (Texas, Seattle and NY *ALL* beat their pythag in 2009, fyi). Not saying this is conclusive evidence.
I'm not denying that club has to score more than it allows. It does.
My issue is the notion that - "we can sacrifice a little defense to get some offense" actually does NOT produce more wins. That is precisely what happened with Tampa in 2009.
If you sacrifice a little defense for a MAJOR jump in offense, then you gain ground. But, the numbers and trends seem to indicate that most teams (that I can find), that swap defense in exchange for offense typically fare WORSE for the exchange.
For the most part, the sabr world understands that it doesn't understand defense. It's still in the "here's our best guess" stage of growth. But, just about every model around assumes that pitching is 50% (or more) of the run prevention equation. But, if that's true - how could a 9th-best PITCHING squad end up on a team #1 (by a wide margin) in run prevention?!?
If pitching is equal (or larger) in importance than defense - then 9th + 1st should equal 4th place (or worse). *IF* all of the current models which hand 50% (or more) of run suppression to pitching are wrong, (and I believe they could be),
In '09, Seattle was first in run prevention with the 9th best pitching, plus the best defense. What if they retained the best defense and moved to 1st in pitching? How many runs do they allow in THAT situation? 600? 575? 550?. How many wins for a team scoring 640 and allowing 550? (Just fyi - that's almost exactly what the Braves did in the strike-shortened 1995 season in which they won the WS).
Personally, I LIKE the fact that Z has worked harder to tweak the bullpen than the lineup. Lee and League are huge. If Bedard does return in May ... and RRS continues to develop ... and ANY of the other spaghetti, (Snell, Fister, Hill) puts it together, then if the DER doesn't slip I believe that yes -- it IS possible for the club to score 650 and make the playoffs ... because they'll only be allowing 550.
Sandy...thos emid-90s Braves are a good example of the closest a modern team has ever come to assembling a defensive juggernaut model. It bears repeating that the 1969-1973 Orioles were also in that style...with very similar problems on offense (Mark Belanger played every day and struggled to hit an empty .240. :) ).
I do think my nuymerical evidence is solid that pitching is about 55% of run prevention...not less as you suggest...however, I am on the extreme low end of the sabermetric envelope for pitching vs. defense breakdown and my defensive analysis gives a lot more weight to fielding than WAR or anything else out there.
The Mariners of 2010 are a great test of my theory Because their pitching is only going to be plus plus if they get 120+ innings out of Bedard and 150+ out of RRS. Any other outcome relies very heavily on fielding.
The value of hitting in a ballgame is equal to the value of pitching plus defense. Offense is the most important of the three.
Pitching and defense seem to be interrelated... you can't value them separately, because they affect one another so strongly. I think that to say 20 out of 27 outs (or whatever) are entirely defense and not at all pitching... is vastly overstating the situation. However, I can't tell you if the sliding scale is linear, exponential, or something else.
What I CAN tell you is this: if your defense is AWESOME... pouring a high percentage of your team's money into pitching is less necessary. And if your pitching is AWESOME... you get less benefit from your outstanding defenders.
If Cliff Lee gets 5K more per game than our #5 starter, then on those days an infield of Kotchman/Wilson/Figgins and an outfield of Ichiro/Gutierrez gets 4-5 less chances to shine, and their defense becomes proportionally less valuable. The same would be true of any pitcher with a lower contact rate, I suppose.
I think that a team needs to build up both sides of the pitching/defense half of the game, but not both to the extreme at the expense of offense. It seems possible that this could be an inefficiency in use of resources.
---
Also - diminishing returns in defense will happen, if nowhere else, when two excellent fielders' effective range overlaps. It will also happen if you examine a 2B who errs on 10% of his throws to first and a 1B who errs when receiving 30% of bad throws to first. If you replace the 2B with a Gold Glover who errs 5% of the time or the 1B with one who only loses 15% of bad throws... you cut errors in half, and gain back half the runs lost on defense in that transition. If you get BOTH, however... you don't get the full expression of both of their skills. The 1B has only 50% of the marginal 'bad throws' to redeem, so his specific subskill is only half as valuable.
Usual excellent thought-provoking stuff, Sandy & crew!
Baker reports that some of today's workout had Lopez at 3b and Figgins at 2b.
Other than having Figgins ready to play second if there were a Lopez trade, is there much to be gained from flipping them? Would it make a difference if Tui ends up on the bench as a backup at 2b-3b?
There is a balance out there somewhere between pitching, defense and hitting. We have lots of examples of tipping the scales to offense, sometimes with success, often without. Tipping the scale to defense is revolutionary (retro maybe? Braves, Orioles) by today's models. Z is breaking some new ground. He already has the big east teams checking their D, quote Theo: "We were focused on building a balanced club,'' GM Theo Epstein says. "Our goal was try to be above average in hitting, pitching and defense. The years we've been able to accomplish that have been our best years -- '04,'07 and '08.'' (Smart guys know a good idea when it's worth swiping.)I'm excited to see where Z might find the balance that tips the scales in the M's direction. I don't think anybody called the 24 games bump last year. We're covering new territory here.
What King said - I'd spoint that if I could.
Expanding the potential fielding zones out to their utmost at each position is good in theory, but at that point there should be significant overlap with other fielders. Is it necessary to do so? Is it more beneficial than adding hitting, which by its very nature cannot overlap in that fashion? The marginal returns on multiple fielders would seem to be in the amount of excess fielding gained (before overlap) when compared to the drop of linear production from putting a bigger bat out there.
Everyone was fond of the 3 CF phenomenon of running Cammy, Ichiro and Winn out there to help the pitchers. But to what extent could we have put an Adam Dunn in LF, let 2 CFs cover more of each of their potential zones, and upgraded a 99 OPS+ bat for a 132 OPS+ one? How badly would that have affected our staff vs. our offensive production? If the D for Dunn's section of the field drops 15% (with Cammy shouldering extra duty in Dunn's zone, as he was wont to do into Winn's zone anyway) but his positional offensive impact goes up 30%, how is that a loss? DER would go down, but at what point is that an acceptable tradeoff? It has to be acceptable at SOME point. What is it?
I don't know what the figures would be. But I also believe that defense is not the sole reason the Ms improved by 24 games and the Rangers by 8, and that Tampa did not lose 13 games just by going from #1 to #4 in DER.
It was certainly part of the equation. But if we're talking "M's obsession with defense" vs. "M's preference for defense" then the percentage gain we got from the D and the amount of glove that can be sacrificed (probably at specific corner positions, as has always been known to baseball) is what we're talking about.
Ryan Braun was a -6 in fielding RAA, right? Franklin Gutierrez was almost 29. Ichiro was 19.
At what point can I play Ryan Braun in LF and let the other gloves make up the difference, instead of trying to find another plus-plus fielder so my team DER doesn't drop from #1 to #4 and send me plummeting out of the playoffs as you intimate happened to the Rays?
It seems like teams that go looking for offense don't really understand defense, but also that defense is immensely volatile as a stat. It's not like runs scored change as much as defense, where the top 3 teams from this year in the AL were the bottom 3 teams from last year.
So how does a team maintain its defense at a high level while looking for more offense?
~G
Defense is important, about 15-17% of the game of baseball give or take, and impact defenders can win you a pennant. That's been true not just since Belanger, Robinson and Blair, but since John McGraw and before.
Questioning Jack Hannahan at 1B isn't questioning the value of defense as such. :- )
Or Kotchman at 1B.
Thanks for the compliment, G - I count that high praise.
This is exactly what I'm talking about. If you start with the "obvious" -- runs scored and runs allowed are each 50% of the baseball equation -- then if you split pitching/defense straight down the middle - defense would be 25% of the game. Yet Doc uses a pretty standard (and generally accepted in some quarters) standard that defense is only 15-17% of the game, meaning pitching is 33-35% of the equation.
My point is that this number is drawn from a bunch of different approaches which all, (in some way or another) DISCARD a bunch of defensive events, while counting every pitching event.
It is completely indisputable that fielding gets 20 outs and pitchers get 7. That's not speculation. That's fact. And ALL of these individual defensive metric methods start off by thinking of every way they can to ignore some portion of those 20 outs the fielders accumulate each game.
Each year in the AL, the 1st to worst spread in hits allowed is about 300. The runs allowed differential is eerily similar (roughly 300). Pirato ran a correlation and found that the hits to runs slope inside the standard pitching paradigm of the AL is effectively 0.8. Each hit prevented prevents 8/10 of a run. The thing is - it's actually not linear - it's hyperbolic, so from end to end, it skews toward 1 hit saves 1 run. But, most of the defensive metrics use much smaller run-per-out factors.
During the 20th century, run prevention was assumed to be nearly all pitching, yes. And the only defensive stat for most of that century was fielding % ... errors. The difference between your SS making 20 errors and 10 errors was largely what they had to go on.
Clearly, the baseball world remains in the stone-age in regards to defense. Felix finishes 2nd in the CY voting because his ERA dropped to 2.49 (from 3.45 in 2008). But, his personal stats were largely that same as his 2006 season, (K and BB rates). Washburn had an ERA with Seattle of 2.64. But, the club can't buy a GG outside of Ichiro.
The Braves won 15 straight division titles -- and they were top 3 in DER in something like 12 of those seasons. They ALSO had top 3 pitching in most of those seasons. And they also had middle-of-the-pack offense in most of those seasons.
Agreed with most of this.
I would like to carefulyl clear my throat and say that I do not discard fielding events more than pitching events...both of them have a standard marginal discard pile and the size of the margin is the same for both. That's probalby why PCA has always given more fielding credit than any other sabermetric tool.
We need an outbound radar gun on balls in play that captures the vast majority of game action. That begins to close the knowledge gap on how much influence pitchers have over BIP, and also how much value defense brings to the table. A GB pitcher who gives up 110mph smokers but has a GB% of 60 looks BETTER (using today's information) than a GB pitcher who gives up 80mph bleeders at the rate of 55%. Same goes for flyballers. If a guy gives up 120mph cannon shots to the OF 35% of the time, is that more valuable than a guy whose average velo going to the OF is 105mph 45% of the time?
We have PitchF/X. We need outbound velocity now. If we get accurate information on how HARD hitters are smacking the ball around, it will begin to clarify the defensive debate.
The Braves had THREE HOF starting pitchers, too, and a ton of other good arms. Or were they only HOF pitchers because DER obfuscated their weaknesses?
I don't have any problem giving credit to defense (I love leather...), but unless you're engaging in serious hyperbole for the sake of advancing discussion on the topic, it seems to be your argument that the Ms could have increased their 2008 win total by 24 wins as well just by playing better defense. That the new makeup of the team has nothing to do with it, and Felix didn't change anything to help out - the ONLY variable was DER. Last to first = 24 wins. Terrible to top 3 = World Series for the Yankees. Rangers get better with the gloves and boom! Winning team.
The 2008 Yankees team allowed 727 runs (and scored 789).
The 2009 Yankees team allowed 753 runs (and scored 915):
The Yankees didn't win the title solely because of defense. And most of the defenders were the same - though Texeira replacing Giambi is obviously a huge increase. And having a better 1B should make the rest of the IF better, though the individual (poorly measured?) numbers don't bear that out. I still think the extra 126 runs scored helped more than Teix's defense at first. Call me crazy.
Texas, 2008: Scored 901, allowed 967
Texas 2009: Scored 784, allowed 740.
Man, if they'd scored what they had in 2008 they coulda gone all the way...but yes, shaving TWO HUNDRED AND THIRTY SEVEN runs-allowed off their 2008 figure made for a significant improvement. I didn't need DER to tell me that. Was that ALL defense? If it was then defense is the most volatile part of the game, because that swing can happen with most of the same contributors, in the same positions, with the same coaches. How do you break out which parts were defense, which were pitching, which were luck...?
Sabermetrics poo poos the measurability of chemistry, but defense would certainly seem to involve some "chemistry" and some "gelling" as well as hated luck and the intangibles of coaching that we can't put a finger on statistically. All we know is what happened, not WHY it happened. How can GMs try to add this defense thing that you claim is so under-rated if they don't know what makes it work in the first place?
Tampa Bay, 2008: Scored 774, allowed 671
Tampa Bay, 2009: Scored 803, allowed 754
So they improved their scoring by 29 and dropped 83 runs in the allowed column. 13 games seems like a harsh drop for that, but they also seemed to do it in "unlucky" spots. These things happen, and they should have dropped wins. It wasn't that they made a mistake choosing offense over defense, it's that they didn't add ENOUGH offense.
The Yankees gave up more runs this year than last year - they just scored a TON more. If the Rays had added 150 more runs instead of 30 they'd have made the playoffs just fine and we wouldn't be talking about their DER drop. The more likely answer to "why did they give up on defense?" is that they DIDN'T - 754 runs allowed isn't a lot. They just knew they were lucky last year to get as great a defensive performance as they did and tried to add offense to make up for the rebound back to their gauge of normal, and couldn't quite pull it off. Their GM is no longer stupid - if your decrease in runs allowed isn't maintainable then you'd better add a few more to the ledger in your favor.
Seattle, 2008: Scored 671, allowed 811
Seattle, 2009: Scored 640, allowed 692
So we improved our runs allowed by 119, but lost 31 on offense. That's a mighty big jump for that differential, but we won a lot of 1-run games and a decent number of low-scoring games. We had a year like that before, just a couple of years ago where we outperformed numbers, tho that time with lucky offense keeping it close with bad run prevention instead of the other way around.
Bavasi apparently knew it was a lucky year and knew it could cost him his job if it went the wrong way so he made a huge trade to bolster the pitching and try to maintain the run differential when his offense crashed. Guess what Jack did this year? Our 1-run record and "lucky D" might not be sustainable, so Jack bought the best pitcher he could find via trade to try to suppress runs and help his defensive numbers.
Will it work? We'll find out. I still think we're short on offense. Franklin Gutierrez comes back to earth in his defensive numbers and where does that put us? They say defense never takes a day off so it's more bankable than hitting, which is a streaky side of production...but I don't believe that. Not being able to properly measure defense does not eliminate it's potential streakiness.
*shrugs* But I can't tell what you're trying to tell me, Sandy. Defense doesn't get enough credit? Could be. The Ms could build a long-term contender around defense and pitching and "enough" offense? Again, they could - if they can sustain both the pitching and the defense, as the Braves did. A great D can made adequate pitcher look good, and may have contributed to the reputation of the pitching staff down in Atlanta unnecessarily. Do we know how to keep the defense great? We'll find out.
But beyond that, what do you want me to know? Run differential wins games, better defenses have a better shot in low-scoring games...but we still don't have any clue at what point a lot of defensive emphasis becomes too much which is what Doc's original comment was about. You seem to suggest that their is no point of diminishing returns, but I can't see how to prove it either way except in theory. Is defense an individual thing or a group effort where everyone's performance increases everyone else's? To what extent can a great defense calm down pitchers who know that their mistakes will be vaccumed up with no damage caused?
Defense is mostly theory. It's not that baseball WAS in the dark about defensive metrics - they ARE. WE are. You disagree when Doc puts a vague number on something that you yourself state you don't trust individual metrics on at all. And the group metrics will tell you what team had great defense, but not WHY, and not what moving parts can be safely sacrificed while maintaining defensive integrity.
GMs HAVE to theorize on those moving parts. They can't be paralyzed trying to maintain the defense and unable to upgrade the team for fear of messing up a part of the game that can't be quantified yet. Is the tendency to go with the known (hitting contributions) over the unknown (defensive ones)? It sort of has to be, doesn't it?
So what's the problem, and how do you fix it?
And man is this getting long, so I'm gonna shut up and have another glass of wine. :) Hopefully it'll all make sense in the morning - otherwise I apologize in advance.
~G
The old Orioles with Brooks, Belanger, Blair, and Boog were some of my favorite teams after Bud stole the Pilots. At least the first of those three were big defensive contributors. For the 69-73 teams, all but the last didn't have the DH, so were essentially comparable to NL teams, at least in terms of pitchers batting. Usually, they only had 1-2 batters, including Belanger, with OPS under .700, plus the pitcher. They didn't exactly have difficulty scoring runs, finishing in the top 3 in the league in all but one year. Having a few offensive black holes won't kill your offense. Having four (plus a bad DH) like the 2009 M's will. The A's and Royals had similar team OPS to the M's. The A's with only two positions under .700 OPS scored 119 more runs. The Royals also with four black holes on offense still outscored the M's by 46, but were easily next to last in RS. I expect the M's to show better team OPS in 2010 and score much closer to league avg, just by replacing some black holes with better hitters. C and SS may still likely be under .700 OPS, but LF and 3B should be vastly improved. Also, I sure hope we get more out of DH, since there is no benefit playing a gold glover there. Kidding, sort of.
It is completely indisputable that fielding gets 20 outs and pitchers get 7. That's not speculation. That's fact. And ALL of these individual defensive metric methods start off by thinking of every way they can to ignore some portion of those 20 outs the fielders accumulate each game.
This is the part I disagree with. YOU are assuming that the pitcher has ZERO impact on the 20 outs you credit to fielders.
Just because a fielder ends up with the ball in his glove does not mean he deserves all of the credit for the out. Before you can assume that, we have to talk about groundball pitchers, pitching matchups, intimidation factors (Randy Johnson, much?), game situation, and much more.
I'm not qualified to have that discussion - I'll acknowledge that. But I AM qualified to note its existence.
that 100% of the credit goes to the fielder for those ~20 outs. But there simply is *zero* denying that the fielder participates in those ~3/4 of all defensive plays, meaning they have attributable value to each and every one.
We (meaning the generic masses) don't have the information available to dissect the fielding vs. pitching contributions to overall team defense. Sandy is (in an unusual, but not unwarranted display) using a little bit of hyperbole to illustrate a point that gets ignored all too often in the sabr circles, that point being that there is an incredibly strong correlation to team defensive improvements and overall team record improvements, and it appears to be a much stronger correlation than what we generally believe for offense.
Still, he didn't seem to acknowledge that the pitcher impacts the BIP 'overall fieldability,' which you did and I've attempted to do a few times in the past. It's an important, even critical part of the picture that needs to be included in the discussion if we're actually trying to nail down the contributions split between fielders and pitchers.
The fieldability issue is one I didn't touch on, primarily because Voros' McCracken basically determined that once a ball leaves the bat, pitchers have "statistically" little control over it, (with the exception of the HR ... though HRs are also heavily influenced by parks). The actual BIP results of pitchers, regardless of type, end up being determined almost entirely by the defense behind them.
Any pitchers on a given team will EN MASSE tend to fare better or worse (in terms of BIPA) as a group.
Mind you - I *DO* believe that there are a few (very few) pitchers who slip into that gray area where they make balls "easier" to field. Washburn is my own #1 on that list among active pitchers. But, even these guys lose out HUGE compared to defenses significantly skewed from average. You stick Washburn in front of a bad defense, he sucks. You stick him in front of a great defense, he's great. The skew in the defensive results routinely overwhelms the difference in pitching-specific stats.
Glavine is a sure-fire Hall of Famer. Glavine, while throwing in the lesser league, without DHes, has an IDENTICAL K/9 -- an inferior BB/9 -- and only beats out Washburn in a single stat -- HR/9. Per 162 games, Wash surrendered 27 HRs, while Glavine allowed 18 (for their careers). Washburn allows 9 extra HRs, but walks 12 fewer hitters per 162 games. One is a sure-fire HoFer ... the other isn't even remotely in the discussion.
Glavine had top 3 DER teams behind him for nearly 15 years. Washburn had "elite" defenses behind him just a few times, (2002 and 2009 we had #1 team DERs by wide margins).
The thing is "most" defenses are average. With an "average" defense, then results are going to be determined almost 100% by the strength of the pitcher. But, with average pitchers, results are determined almost 100% by the strength of the defense.
Yes, the two ARE related. Bad/good defenses will "induce" pitchers to alter their patterns. A "bad" defense will typically increase walks, (as pitchers feel they need to nibble more). A good defense may increase Ks ... as pitchers are more aggressive, and pitch ahead in the count, (and with bases empty), more often.
But, the standing principle of pitcher/defense analysis is to "assume" pitchers deserve the bulk of the credit for results. MY belief is that the idea that pitching is 50/50 or 60/40 or 20/80 is a flawed premise from the get-go. At the edges, the "credit/blame" can skew 90/10 to EITHER side. If you DO have an average defense, then results are 100% pitching. If you DO have an exactly average pitching staff, then 100% of the difference from average is due to the defense.
Yes ... I do believe that having a consistently plus-plus defense turned a pretty good pitcher in Glavine into a sure-fire Hall of Famer. And I believe if Washburn had played in front of as many top-end defenses in his career, his typical season might easily have been 15-10 instead of 12-12 ... and the questions would be, if Washburn hadn't taken so long to get going, (first 100 IP season was age 26), how he might have an outside shot at the Hall.
Is Denny Neagle heading to the hall? He went 20-5 and 16-11 during the '97/'98 seasons with Atlanta. The ERAs that season for the Braves starters? 2.20; 2.96; 2.97; 3.02 (Smoltz was the worst).
Millwood had the unfortunate timing to reach Atlanta just before their team DER began to fall apart. John Burkett couldn't get his ERA under 5.50 his last two years in Texas. With Atlanta, in 2001, it was 3.04. His K/BB numbers didn't change much. But his H/9 went from 10.6 (and up) all the way down to 7.7. Oh, facing pitchers almost certainly helped. But, the league skew from AL to NL is about .3 runs ... not 2.5.