Wednesday, January 05, 2005

Ijon Tichy

Last March, Dayn Perry wrote an excellent article arguing that the Dodgers' ballpark should play a major role in its personnel decisions. Perry noted that Dodger Stadium was pretty much neutral for walks and home runs but hurt singles and doubles and triples especially. From this, he drew the conclusion that the Dodgers should seek out "Three True Outcomes" hitters, and, as I noted on Friday, that's pretty much what they've been doing. I'll add here that Jeff Kent does not fit that paradigm, but that's kind of the point: Jeff Kent was signed because he came at a clear discount and because his value in other areas was relatively undervalued, and I'll also note that for purposes of lineup construction having a player like Kent in an order otherwise composed of TTO guys can provide value maximizing power. But that's a story for some other time.

When Perry wrote the article, the Dodgers were coming off of a season in which their run prevention was absolutely dominant but their run scoring was thoroughly wretched. Since then, DePodesta has completely overhauled the offensive, making very clear improvements, but the pitching staff, if we are to believe the clamor in the media, is in a state of total disrepair. DePodesta's job, right now, is to pick up pitchers. Logically, if park factors should play a role in what type of hitters the Dodgers should acquire, they should also play a big role in what type of pitchers they acquire.

Another Baseball Prospectus author, Jay Jaffe, referenced the impact of the Dodgers ballpark on the types of pitchers it should get on Monday. Jaffe casually noted that Javier Vazquez, a fly-ball pitcher, wouldn't benefit from Dodger Stadium while elsewhere implying that ground-ballers Mike Koplove and Derek Lowe would reap Chavez Ravine's benefits. I love Jay's work, but I think he's got this one backward.

You're already familiar, I'm sure, that the Dodgers had excellent defense in 2004. Defensive Efficiency Ratio shows the Dodgers clearly in front for 2004, even when park-adjusted (BP subscription required). But not all balls in play are created equally. One thing I've long been aware of, thanks to the stats at The Hardball Times, is that the Dodgers had the lowest Groundball/Fly ball ratio in baseball last season. Their G/F of 1.05 was significantly below the 1.23 league average. Fly balls are more likely to be converted into outs than ground balls, meaning that some of the Dodgers high DER should be a reflection of the type of balls they're responsible for. Fly balls hits, however, are for extra bases much more often than ground ball hits, as Mitchel Lichtman has found:

fly ball pitchers, on the average, will have a different $H [batting average on balls in play] than will ground ball pitchers, since a fly ball has a higher out percentage than a ground ball. In fact, extreme ground ball pitchers have a BABIP of .297 (1992-2003), whereas extreme fly ball pitchers have a BABIP of .281 (extreme = top and bottom 10% in G/F ratio for pitchers with at least 100 BIP in a season). Of course, the run value of a FB hit is greater than that of a GB hit, such that the actual run value of all pitchers BABIP is almost exactly the same, regardless of their G/F ratios.

So if two teams have the same defensive efficiency ratio - the number of balls in play converted into outs - they won't necessarily have the same run value for balls in play, and the tendency would be for the team with more fly balls to give up more runs.

So let's put this to the test. I collected all the pitching data for NL teams and looked at how many singles, doubles, triples, ROE's (reached on error), and outs they collected. I didn't have sacrifice hit data, since for some reason no team pitching stats page I found tracked that and I didn't feel like totalling all of those up by hand. Then I used linear weights to determine how many runs each team gave up. This wasn't quite ideal, as the lwts figures I had include ROE's as outs since they are used to evaluate offense, but the difference overall should be minuscule if I weigh ROE's as if they were singles. Doing that, I got linear weights run values for each team's balls in play, which I then normalized to the league average to come up with actual linear weights runs saved. The Dodgers did excellently, saving more than ten runs better than the next team, the Cardinals. The defense of the Dodgers and Cardinals saved 80.6 and 70.1 runs, respectively, with this measure, much better than third-place Florida (26.2 runs).

But how do these standings compare to DER? Well, I also calculated each team's net defensive outs, or (Team DER - league DER)*(Balls in play). Fortunately, the scale for outs and runs relative to league average in this context is virtually identical, so comparison is easy. Here are the final numbers, along with the net singles, doubles, ROE's, etc. for each team:

ROE% nROE 1B% n1B 2B% n2B 3B% n3B LD% G/F IF/Fly $r netO diff
COL 0.008 23.3 0.229 -54.9 0.073 -26 0.009 -9.1 0.19 1.27 0.135 -65.6 -66.6 -1
CIN 0.015 -6 0.208 42.7 0.081 -64.7 0.007 1.7 0.192 1.14 1.14 -39.2 -26.3 12.9
ARI 0.021 -31.7 0.221 -16.3 0.067 1.1 0.009 -8.3 0.194 1.27 0.121 -48 -55.1 -7.2
PHI 0.01 15.1 0.206 46.1 0.071 -19.1 0.008 -3.4 0.181 1.09 0.137 23 38.8 15.8
SF 0.014 -1.7 0.212 22.4 0.068 -5.1 0.01 -13.7 0.183 1.24 0.125 -7.9 1.9 9.8
MON 0.013 1.7 0.22 -13 0.064 12.7 0.007 3.4 0.189 1.24 0.114 9.5 4.7 -4.8
MIL 0.019 -23.4 0.208 37.2 0.074 -28.9 0.007 1.3 0.179 1.17 0.127 -18.7 -13.7 5
PIT 0.012 5.4 0.225 -34.2 0.068 -4.6 0.008 -2.3 0.188 1.26 0.123 -30.8 -35.6 -4.8
NYM 0.019 -26.5 0.216 4.5 0.063 17.8 0.005 9.4 0.188 1.29 0.135 14.8 5.2 -9.6
SD 0.015 -7.6 0.212 22.4 0.07 -12.7 0.008 -3.7 0.19 1.14 0.117 -7.3 -1.7 5.6
FLO 0.008 21.5 0.214 11.5 0.066 6 0.008 -4.8 0.182 1.17 0.137 26.2 34.3 8.1
HOU 0.014 -1 0.223 -23.8 0.068 -4.5 0.008 -4.6 0.195 1.24 0.128 -30.7 -33.9 -3.1
LA 0.007 29.5 0.214 13.1 0.059 35.8 0.006 5.8 0.19 1.05 0.147 80.6 84.2 3.6
ATL 0.016 -9.4 0.235 -80.7 0.056 51.1 0.004 15.9 0.193 1.42 0.1 6.1 -23.2 -29.3
CHC 0.013 3.7 0.222 -23.4 0.059 31.8 0.008 -0.8 0.191 1.27 0.137 18 11.2 -6.7
STL 0.012 7.2 0.206 46.2 0.065 9.3 0.004 13.1 0.181 1.45 0.112 70.1 75.8 5.6
tot 0.013 0 0.217 0 0.067 0 0.007 0 0.188 1.23 0.127 0 0 0

"$r" is the lwts run amount, and netO is the net outs amount. "diff" is the difference between the two. You'll also notice the LD%, G/F, and IF/Fly columns, taken from THT (have you bought your THT annual yet? You really should; it's a great book and they should be getting some compensation for the great services their site provides). Now here's what should stand out: all the teams with a G/F below average had a higher net outs than net lwts runs and almost all the teams with a G/F above average had higher net lwts runs than net outs. The only exceptions were San Francisco and St. Louis. St. Louis seems easy to explain: they probably had the best infield but their outfield was weak, so their net reduction in XBH was probably more from the lack of GB than from defense. San Francisco can be similarly explained, as they had very poor outfield defense and played in a park that favors doubles and triples.

As you can see, the Dodgers gave up more lwts runs per out than average, as their G/F tendencies would predict. However, the difference was pretty negligible; the absolute value of their difference was the smallest of any team but the Rockies even while their difference from the G/F average was the second most extreme. You might intuitively expect the Dodgers and their FB staff to skimp on singles and be average with the doubles, but that's not how it worked. They were solid across the board. They were excellent in terms of ROE, good in terms of singles, excellent in terms of doubles and good in terms of triples. Contrast this to the other FB-heavy staffs: SD, PHI, and CIN all gave up a bunch of doubles while clamping down big time on singles - each of them saved more singles than the Dodgers.

Remember what I said about the Giants in explaining the reason they differed from the norm? It's the ballpark. The Dodgers gave up few doubles for a reason: Dodger Stadium hates doubles with a passion, and it's not fond of triples, either. Looking at ESPN's one-year Park Factors, Dodger Stadium was the worst venue for doubles in 2004; it's the anti-Fenway. Dodger Stadium keeps hits down in general, but not by as much as you would think. In fact, if you look at the Dodgers' league-leading IF/Fly ratio, the little evidence I have suggests that most of Dodger Stadium's hit reduction prowess lies in its abundance of foul ground. So while the Dodgers did, without question, have excellent defense last season, I think it's clear that a good chunk of their defensive success can be owed to the fact that they benefited from the rewards of fly ball pitching (cutting down on singles and ROE's) while playing in a stadium that eliminated the balancing risk of fly ball pitching (extra base hits).

Now, you might be thinking that this finding is counter-intuitive since the Dodgers had, by all accounts, outstanding infield defense last season. Agreed. The Dodgers terrific infield was, obviously, a tremendous part of their defensive success. One would expect that group to account for much more than 13 net singles saved. However, the Dodgers also had two of the worst defenders in baseball in the outfield - Shawn Green and Steve Finley - playing a solid chunk of the time, and while their other outfielders were all good (well, probably not Grabowski) Roberts was oft-injured and playing a position he wasn't familiar with and Bradley was similarly unfamiliar with left and right field while playing there late in the season. In fact, in September, with Finley in tow and Green playing several games in the outfield, the Dodgers DIPS wasn't far off what it was the rest of the season (given their competition and excursion to Colorado, that shouldn't surprise), but their ERC and ERA were both much higher, even though on the season their ERC and ERA were much lower than their DIPS. Is that the sample size police at my door? Anyway, factor in that Cora and Green were both poor defenders range-wise, and I think it all adds up to a solid explanation.

So, to clarify: the unique features of Dodger Stadium make it ideally suited to a fly-ball pitching staff. A little of that advantage may be ceded with the current renovation which will reduce the amount of foul ground, thus eliminating some of the pop fouls that have held down hits. However, I see no reason for the XBH inhibiting powers of Dodger Stadium to decline from the renovation, so going forward it's clear to me that the Dodgers should value FB pitchers more highly than GB pitchers.

Not to get all Ijon Tichy on you, but this wields a lot of explanatory power. I think this certainly supports the notion that Kaz Ishii can rebound, as I discussed earlier. He's adjusted his approach to get more fly balls, and that strategy can yield great benefits with good outfield defense, especially if he learns that he should still try to strike out southpaws, something he neglected to do last season. Furthermore, it would suggest to me that pitchers like Derek Lowe and (gulp) Odalis Perez are not well-suited to pitching with the Dodgers, and pitchers like Javier Vazquez and Kevin Millwood would be well-suited to LA. Now, you might think now: but if Dodger Stadium's neutral on home runs, fly balls pitchers will still give up a lot of home runs. Well, that depends on the pitcher; it's for that very reason that DIPS and FIP (metrics of defense/fielding independent defense) both include home run measures. Obviously, I'm not saying that given two pitchers the Dodgers should choose the one who allows more fly balls; when choosing pitchers, the Dodgers should, of two pitchers with roughly equal DIPS projections, choose the one that is more of a fly ball pitcher, all else being equal.

I've expressed, on numerous occasions, my feeling that Odalis Perez is not hot stuff, despite the flashy ERA's. Certainly, though he's not an extreme ground ball pitcher, the relative value of FB pitchers for LA hurts his case. He's actually had his most success with the Dodgers in years when he gave up more fly balls:

2002: 1.36 G/F, 3.00 ERA, 3.54 FIP, 2.34 ERC, 17.8 K%, 4.4 BB%, 2.4 HR%
2003: 1.99 G/F, 4.52 ERA, 4.39 FIP, 4.11 ERC, 18.3 K%, 6.0 BB%, 3.6 HR%
2004: 1.62 G/F, 3.25 ERA, 4.29 FIP, 3.29 ERC, 16.3 K%, 5.6 BB%, 3.3 HR%

Let's not go crazy with that, though, as he's kind of all over the place. He doesn't allow many line drives, so I'd expect his ERA in front of a neutral defense to be in the 4.00-4.10 range. The Dodgers will have excellent infield defense next season, perhaps better than what they had last year, and they'll have excellent outfield defense too. All told, I'd be surprised if Perez' ERA was higher than 3.80-3.90 or so, but I'd also be surprised if it was lower than the 3.25 of last year. Wow, I've just established a pretty slim margin of errorfor myself, haven't I? I'm not sure how much I like the Perez deal, as he's overrated (except by people who only look at W-L) but the Dodgers clearly did have a need to pick up someone, and in the current market $8 million per isn't a terrible deal for him, and I doubt paying a million per extra for Matt Clement would have been worth it. That being said, though I'm certain I know less about the injury status of the players involved than Paul DePodesta or others, from a performance analysis standpoint I think Kevin Millwood figures to be a better deal. Of course, Millwood's still out there, and the Dodgers do have enough cash left to get him, so who knows?

Furthermore, this fly ball business touches on that other Dodgers offseason bugaboo, J.D. Drew vs. Adrian Beltre. Simply speaking, if the Dodgers maximize the utility of their ballpark by acquiring fly ball pitchers - which, to an extent, they have - then they accentuate the significance of their outfield defense. UZR indicates Shawn Green's right field defense is out 20 runs below average in a neutral context; with a heavy fly ball staff, that could translate into a value of nearly 30 runs below average, or 3 wins. Conversely, J.D. Drew is in the +10 range, which means the swing between the two defensively could easily be exceed 4 wins with a FB staff. Now, J.D. Drew also appears to come with the cost of moving him to centerfield to ease his knees. If that's the case, he looks to be a roughly average centerfielder, and some of Milton Bradley's value is lost with the move to right field. However, I would expect that to be only about a 10-15 run swing, so overall we're still looking at a 3-run defensive improvement versus Green. You can talk until you're blue in the face about Adrian Beltre's defense; he's great, and he's a +20 win defender. But third base is smack dab between RF and CF on the defensive spectrum, and check out the major league positional averages for each position:

RF: .272/.348/.449
3B: .273/.340/.455
CF: .272/.335/.437

In a context-neutral sense, Beltre doesn't have any major edge over Drew based on their 2004 performance levels. Plus, while both Drew and Beltre had career seasons, Beltre's got a whole lot more regression to the mean potential than Drew. That is to say, Beltre's season was at the upper bound of offensive performance, and it's extremely unlikely that he can improve on it; if he declines, there's a long history of him performing at a much lower level, whereas Drew has established a performance level substantially north of Beltre's. So while it's true that, going strictly by age, Beltre is entering his nominal prime while Drew is already in his, that doesn't really portend that Beltre's performance relative to Drew's will improve. Given all that, I don't really see where Beltre earns $2 million per season more than Drew in a context-neutral sense. Heck, back in September I concluded my ridiculously long Beltre analysis by saying he'd sign for 5 years, $55 million. But in the context of Dodger Stadium, Drew already adds relative value because a value-maximized staff will be fly ball heavy. So independent of whomsoever else is involved, I think Drew is a better value than Beltre at 5 years, $55 million and it just so happens that the Dodgers would need to pay another $9 million for Beltre.

But, of course, there are others involved, in the persons of Shawn Green and Jose Valentin. Green is a better offensive player than Valentin, although the White Sox absurd decision to have Valentin start against southpaws last season makes his offense look worse than it really is. Valentin, if wisely deployed is about average offensively, and Dayn Perry would probably agree that Valentin's TTO prowess makes him a good fit for the Dodgers. Maybe you're really into that super-popular stat that Green was .253/.335/.399 before the all-star break and .281/.371/.529 after it. A much less popular stat is that in 2003 that same split would be .255/.317/.429 pre and .316/.408/.506 post. I don't see what reason there is to believe that the second half of 2004 represents his true performance level and not just small sample size noise. Given the likelihood of his torn labrum rearing its head again, +20 runs for Green's offense is slightly optimistic. So Green has a 20 run edge on Valentin offensively, but, as we established before, the addition of Drew creates a 30-run swing on defense. Given Valentin's tremendous defense at shortstop, he should match what Beltre did defensively, if not exceed it. So the offense of Beltre and Drew cancel out, the defense at third cancels out, and the outfield defense advantage gives an extra win relative to Green's offensive edge over Valentin. One win gained. If the Dodgers had re-signed Beltre at $13m per and kept Green in place, they'd be on the hook for $29 million in 2004. Now, they're paying $22.5 million to replace them with Valentin and Drew and to rid themselves of Green's contract. So if my arguments are accurate, the Dodgers save $6.5 million to pick up an extra win. They used that $6.5 million on the difference between Alex Cora and Jeff Kent, which probably adds a win or two defensively and two wins offensively. So for the same price, I see a four or five win improvement. On top of that, they get Dioner Navarro and William Juarez and owe Drew $9 mil less over the next five years than they'd owe Beltre. We could also throw into the discussion that the Dodgers have several solid third base prospects and few solid outfield prospects, enhancing Drew's marginal value. But why pile on?

Oh, one last bird to pummel with this particular stone. Derek Lowe. You know him, you love him, you can't get enough of him. Certainly, my fly ball conclusion doesn't help Lowe's case. He's probably better than I mentally give him credit for, and if he were plugged into the Dodgers he'd probably have a sub-4.00 ERA, given his 4.4ish DIPS in the AL the past two seasons and the Dodgers defense. In fact, there's not much separating him from Perez. That being said, Perez provides better marginal value due to his less extreme GB tendencies and Lowe's price tag probably isn't much lower than Perez' - heck, it might be higher.

This hasn't been a perfect offseason for the Dodgers, but it's been much, much better than almost anyone is giving them credit for, and I'm not sure that anyone could have done a better job than DePodesta has done.

Or maybe my computer just has a virus (hey, maybe Simers was inspired by what I wrote here).

>>I don't really see where Beltre earns $2 million per season more than Drew in a context-neutral sense<<

It's called the market, Tom. Quality third basemen are hard to find, much harder to find than corner outfielders, a point you made earlier. I don't understand your surprise at this.
Further: on the one hand, you mention Valentin does nicely against righties, but you don't adjust for the absence of his glove at third against southpaws. 3B in the current lineup will be a platoon of Valentin and Antonio Perez. I don't know if this will make a huge difference, but it strikes me as being nonzero, in any event, and you don't mention it.

>>I don't see what reason there is to believe that the second half of 2004 represents his true performance level and not just small sample size noise.<<

How about, he's six months further away from his labrum surgery? And as to this:

>>Maybe you're really into that super-popular stat that Green was .253/.335/.399 before the all-star break and .281/.371/.529 after it. A much less popular stat is that in 2003 that same split would be .255/.317/.429 pre and .316/.408/.506 post.<<

Green's pre/post All-Star break lines in other years:

Year----Pre All Star----Post All Star
2002 .280/.378/.592 .292/.394/.520
2001 .289/.358/.527 .307/.389/.682
2000 .294/.404/.509 .241/.323/.431

Mr. Small Sample Size rearing his head again?
Gosh darn it Rob, you're starting to get on my nerves. Are third basemen harder to come by than corner oufielders? Of course, there are twice as many corner outfielders. But if you only look at right fielders, as I just showed the data on, the offensive difference is negligible. Drew projects as a slightly better offensive player, Beltre as a slightly different defensive player. The position adjustment is almost neutral. On top of that, I just argued why the situation is not context neutral; Drew has added marginal value, and getting rid of Green while being able to acquire Valentin on the cheap as well as the Dodgers depth at 3B in the minors makes Drew a better deal.

Yes, Antonio Perez isn't as good as Valentin defensively. He'll start forty games. However, he's been a shortstop in the minor leagues, so I imagine he'd be a plus defender at third base. If Valentin is a +25 guy and Perez a +10 guy, that's +22 for the season. That's pretty much what I implied, and if Perez hits mainly RHP he's probably a +3 guy offensively.

As to Green, I wasn't trying to imply that he's a big second half hitter and a bad first half hitter. Rather, I was trying to imply that that second half business had a lot more to do with sample size than a change in talent level. Yes, I understand that being further away from a surgery can increase a player's ability, but he was just as good in April as he was in July, August, or September. I don't think there's any reason to believe that his overall performance will be like that over the course of a season. Even if he has established a new performance level, there's still a substantial chance of regression, and not just toward the mean but also regression below the mean. Any way you slice it, taking his 2004 in full as the baseline is probably the best way to go, and that's what I did.
>>Are third basemen harder to come by than corner oufielders? Of course, there are twice as many corner outfielders. But if you only look at right fielders, as I just showed the data on, the offensive difference is negligible.<<

But... that wasn't an answer to the question you posed. "Context-neutral"? What does that mean? You mean, if they both played the same defensive position? Sure, I'll agree to that, but the question as posed was confusing.

>>If Valentin is a +25 guy and Perez a +10 guy, that's +22 for the season. That's pretty much what I implied<<

Why am I supposed to read between the lines here? I don't get it.

As to Green -- the fact that he'll be playing in a hitter's park (1.290 HR park factor according to ESPN) will be a plus for him. Yes, there's the risk of his labrum going out again, but maybe I'm a little more optimistic on this one than you are. At 32, I can understand the Dodgers wanting to let him go, at any rate.
I guess that wasn't clear. By "context-neutral" I meant a team that could choose either one without it impacting the relative utility of their resources, as in "Which would be better for the average team?" My assessment is that Drew's offense projects as being better than Beltre's by enough of a margin that a) the defensive gap, run-wise, is overcome and b) the negligible position adjustment is overcome.

I didn't mean you should have grasped the A. Perez part from the text of the original post; I meant the value of Valentin, adjusted for Perez, is equal to the value of Valentin that I implied in the original post. I guess my incorrect assumption was that most people could figure out that a shortstop who projects at a little below average offensively would, in a platoon relationship at third base, be just above average. Again, this reflects my writing process: I think as if I were reading it, which is without a doubt the biggest problem with my writing skills.

Hope that clears things up.

And I agree that the BOB should help Green, but then that's true of anyone.
I'm not sure I buy the critical point in your post - that DS favors FB pitcher benefits more than a GB pitcher. Your argument certainly makes sense pre-2005, but with the DRASTIC reduction of foul territory I could see those numbers changing more than you anticipate. Just eyeballing the pics I've seen, it seems that 50% of the foul territory will be gone, which takes away a LARGE advantage to FB pitchers (as you mentioned). Thus, while your analysis is certainly believable pre-2005, I could see it be rendered obsolete next year.
Why does anyone think Jose Valentin is a good fielder? After years as a league leader in errors at his position, he's only improved to barely adequate over the last 2-3 years. The only good thing you could say about his on-the-field skills is that he hits a little better than Royce Clayton (sometimes). As a Chicago White Sox fan, I was glad to see him go (and it should have happened a couple years ago). He is error-prone in the field and he can't hit lefthanded pitching from either side of the plate (he called himself a switch hitter until last spring). But he's handsome and he doesn't spit too much (my wife insisted I say something positive about him).

Charles Koster
San Gabriel, CA
Believe me, I've given a great deal of thought to how the renovation will affect this. The thing is, what I'm isolating here is Dodger Stadium's doubles/triples reduction power, something that, so far as I can tell, is not substantially linked to the amount of foul ground. While fly ball pitchers do tend to induce more foul pop-ups, generally IF/Fly rate should be considered independently of G/F, so that's not really an issue here.

Glad to see my audience isn't just Dodgers fans. I think you're undervaluing Valentin by quite a bit. For one, fielding percentage isn't a very good measurement of fielding proficiency; errors are only a small part of the puzzle, and getting to more balls is much more important than mishandling a few. In fact, players with a lot of errors tend to get to many more balls than their counterparts. The defensive metric that I trust the most is Mitchel Lichtman's Ultimate Zone Rating, which actually uses play by play data which accounts for the location and quality of the balls fielders track down, and adjusts for a ton of things. This data is then compared to league average run expectancy for each type/location of ball, and each player is graded by how many runs above average they are at their position. UZR has rated Valentin as the best shortstop in baseball pretty consistently. I think it makes more sense to use the actual data for defense than to go by what we think our eyes tell us about the player. I hope that answers your question.
>>Given Valentin's tremendous defense at shortstop, he should match what Beltre did defensively, if not exceed it.<<

I am not familiar with the details of UZR, however I would believe it probably parallels the 'traditional' ZR metric to a large degree. Have you looked at Valentin's lifetime ZR at third base? It's abysmal.
>>While fly ball pitchers do tend to induce more foul pop-ups, generally IF/Fly rate should be considered independently of G/F, so that's not really an issue here.<<

Why should they be kept independent?

I don't know... I ran a correlation between G/F and IF/Fly and there was a -.68 correlation, which is not exactly negligible.

also, from THT glossary:

>IF/Fly: The proportion of fly balls that are hit in the infield. Analysis indicates that pitchers do have some control over this type of batted ball.<

So it seems that FB pitchers have a slight ability to induce more IF popups per fly and more fly balls in general, which seems to indicate that FB pitchers greatly benefit from the foul territory in DS.
ZR isn't a terrible metric, but it's not as good as UZR by a long shot. And your contention that his career ZR is bad is, well, false. He certainly had some poor years by that metric in the 90's, and his 2001-2002 ZR's were about average in very limited playing time. However, he's been second in ZR in each of the last two seasons; I don't see how you can discount that.

Dodger Roger,
My point is that IF/Fly should always be considered in evaluating a pitcher. This is independent of G/F ratio. Moreover, my arguments about why Dodger Stadium helps fly ball pitchers operate independently of the amount of foul territory. That is to say, the advantage of fly ball pitchers in getting outs is universal-- fly balls in the field of play will always be converted into outs at a greater rate. What Dodger Stadium affects is the relative value of fly balls since they won't become extra base hits very often. G/F is on balance neutral because the out value of FB is neutralized by its XBH value. This has very little to do with the frequency of infield flies. Given the already extant data on pitchers' ability to induce IF flies, any pitcher evaluation should assign value to a pitcher's IF fly tendencies, independent of G/F. What I'm saying is that, all else (i.e., DIPS, LD%, and infield flies) being neutral, fly ball pitchers are much better at Dodger Stadium. If you have an argument as to how the renovation will substantially alter the park's doubles/triples factor, I'm anxious to hear it. I've spent long stretches of time staring at the photos, and I just don't see it.
I think I understand a bit better what you are saying. I knew that pitcher had some ability to induce IF flys, but I had assumed that it was the FB pitchers that had this ability!! I will have to check this out.

Regarding the changes affecting doubles and triples, I think the geometry of DS funnels all the balls to the fielders, so doubles and triples should only go up a little (since more balls will be put in play with fewer foul outs).
Nevermind about the GB/FB Inf Fly correlation. I did it on the player level and the correlation was close to zero.
hmm... with all the talk about TTO hitters... any rumors about DePo chasing Adam Dunn? How do you think Dunn would fare at Dodger Stadium... any effect on his stats?
One final comment, Tom -- where are you getting your UZR data from? Tom Tippett doesn't have it anymore, neither does Tangotiger.
I thought I sent you the links. The yearly data's not up anymore, but Tango has the weighted 2000-2003 data on his page, and mgl posted incomplete data here:

That rumor's about a year old. The Dodgers don't really have any place to put Dunn-- they've already got three TTO types in the outfield. Plus, Dunn is a pretty poor defender, so he certainly doesn't fit in with the thesis I just presented here. He'd be an upgrade over Werth or Bradley, for sure, but not by anywhere near the amount of talent that would have to be given up to acquire him. Plus, the Reds aren't shopping him at all.
Sorry, I thought maybe you had a complete 2004 set somewhere I was unaware of. Pity.
If you haven't yet, Tom, goto primer and check out this thread:
MGL did an analyis of extreme GB pitchers vs. FB pitchers and came up with the opposite conclusions.

Question: Are GB pitchers more likely to give up doubles than FB pitchers or is it the opposite or is it neither?
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?