I'll bet the Sox could even get Reed Johnson, Juan Rivera, and Raul Ibanez in the right deal.

Now I think Boston would actually be well-served by trading Ortiz for some other help

That's just silly. Unless the Red Sox can acquire Jeter in return, I don't see it happening.

Before David signed his most recent contract, he was the most valuable possible trade commodity in the game. Now it's less so, and NL teams should beware. I'm not suggesting that they trade him just to get him out of town or just to get Manny some DH at-bats, but if I read the perception around the game well, the Sox could solve many of their roster construction problems in a single move that way.

I seriously doubt there are many, if any, NL teams eagerly waiting to sign David Ortiz and install him at 1b. (The usual caveat about teams from Queens, NY, and soon-to-be-DLed ex-Red Sox players applies.)

but if I read the perception around the game well, the Sox could solve many of their roster construction problems in a single move that way.

They'd have a perfect roster but any empty stadium for them to play in front of.

Here are Manny's road numbers. Basically, I used the same methodology I always use, but I removed all data from Fenway.


03 -9 (-22)
04 -5 (-12)
05 -7 (-14)
06 -9 (-22)

So, if you scale these road #'s to 75 games (half of 150) you're looking at:

03 -9 (-22), -11
04 -5 (-12), -6
05 -7 (-14), -7
06 -9 (-22), -11

He's averaging something like a -9 for the road half of 150. I'll buy that at a typical park he'd be about a -18 (or about as bad as left fielders get). Now, to believe that he's really -30 bad, you have to believe that most of his badness comes at home and I'm not sure I buy that. His familiarity with the wall and the smaller area of left field should reduce the damage he can cause (from what I've seen he plays the wall pretty well). I tend to think that if he's -9 on the road, he's more like -5 at Fenway. My WAG (and admittedly, it's throwing out half of the data) is -14 total/150 with half his games at Fenway.

There are about 1.8 plays per game for corner OF'ers.

That's too low.

There has to be more than that. BB-ref stats indicate leftfielders averaged around 2.1 a game for about the last 6 years, and that's just counting putouts and assists. They aren't counting hits that have to be fielded and gotten back to the infield.

I have posted this value on this site before...it is closer to 350 chances, divided by 162 or 2.15 chances per game.

1.865 catches per game in LF for the average LF'er (in 2006). I define a "chance" for a fielder as the number of balls turned into outs that an average fielder would make given the same distribution of balls in play. You (anyone) can define a chance any way you want as long as it is clear what you are using it for and the definition fits the use. When you are using a ZR system, you can define a chance as a "ball in zone." When you are using a system that looks at all balls regardless of where they are hit, there is no logical way to define a chance - technically, every ball is a "chance," especially of the league average catch rate for that ball is greater than zero. You are left with arbitrarily defining "chances" and I choose to use the definition I do. I suppose that a true "chance" would be somewhat higher than the number of balls that an average fielder catches. Maybe the number of balls that the best fielder in the world would catch? Who cares? It doesn't matter. Regardless of what definition you use, as long as you compare a player's "chances" to a league average fielder you can figure out "defensive games" if that is what you are trying to do. And as long as you approximate how many balls a player has a reasonable chance of catching, you can use that to figure standard deviations, standard errors, etc., for a player's UZR (or whatever metric you are computing).

Chris,
you're using chances as defined by STATS ZR, which ignores many line drives. Under a more liberal definition of all the air balls hit at least 200 feet, you'd get more like 500 chances a year. (HBT, now using the new "Dewan" ZR, shows 494 chances per AL team in 2006 because they count all those line drives with a "low" probability of being caught.)

My opinion is that if a system is evaluating the chance that a fielder might make a particular play, that's a "chance." If that's every ball hit anywhere on the field or off the wall, a là Pinto, even though most of them are zero probability for the left fielder, that's no big deal (as long as it's done by a computer instead of by hand!) But of course you have to be fine-grained enough in evaluating that probability, and you also need enough data to make the probability reliable.

For the question of whether to "throw out" the balls off the wall at fenway, in the context of a ZR system, you're much closer to the truth if you do it. But as Darren noted in #27, it likely overcorrects, assuming you then just compared a red sox left fielder to the overall league average ZR. Those deeper balls would be caught at a rate below the overall average across the league in other parks; take them out of the sample and you'd expect a higher ZR. (So there's a further adjustment to figure out for that.) There's an interesting data set available here to illuminate this point. (Look at the bottom of the page under related files for outbyzone.csv) It also gives some idea of the slight jaggedness you'd get with small (10' by ~20') zones, illustrating Mickey's point about that. Naturally that jaggedness is much worse if you construct it for a single park.

Extending what Misterdirt did with retrosheet and including 2006 I get 4 year totals by hit type of:

Fenway
F Manny 363/554 .655 (312 TB allowed; 1.63 per hit) all others 645/948 .680 (501 TB allowed, 1.65 per hit)
L Manny 42/334 .126 (404 TB allowed 1.38 per hit) all others 70/646 .108 (780 TB allowed 1.35 per hit)
P Manny 0/0 all others 1/1 1.000 (0 TB allowed)

Red Sox Away Games
F Manny 370/445 .831 (115 TB allowed 1.53 per hit), all others 774/899 .861 (194 TB allowed, 1.55 per hit)
L Manny 47/365 .129 (414 TB allowed 1.30 per hit), all others 106/694 .153 (747 TB allowed, 1.27 per hit)
P none

If you assumed that the specific difficulty of the batted balls evened out, and if you assumed that these other left fielders (red sox opponents and red sox backups for Manny) were collectively average, and if you assumed retrosheet hit types were consistently judged and recorded without errors, then you could conclude from all this that Manny was -14 plays on flys at fenway and +6 on line drives at fenway, and -13 plays on flys in away games and -9 on line drives, in his actual opportunities. There is no sign that he is better at preventing extra base hits at Fenway, and no sign that he is worse at allowing them away from Fenway. His total bases allowed seems completely comparable in both. Cumulatively over 4 years, he would be -8 plays at Fenway and -22 plays away. Accounting for the hit value of those missing plays, that would be about -10 runs and -23 runs. So that's about -8 runs per year (actual) in an average of 1104 defensive innings. Casting that into the 150G = 1350 inning basis quoted for UZR, you'd get -10 runs a year, (in terms of performance, not talent).

Now that's a lot of assumptions, and to the extent that they are violated, Manny could have been worse (or better) over the last 4 years. There's no obvious reason the assumptions should be expected to break in the worse direction though, which would be necessary to 'confirm' UZR. Fortunately these assumptions can be tested ...

"I gather Griffey would still be a below average defender in RF?"

I would assume. They're still probably better off with him in right and Dunn in left than the other way around, though.

To summarize, here's what we have from each for 06 for runs cost per full season of play:

Arthur: -12 (using an average of his and AROM's park adjustment)
Dial: -12
Chone: -35
MGL: -32

I was the one who averaged my adjustment (14 runs) with Joe's. That -35 runs is completely unadjusted. My projection for Manny is -14 runs.

Before David signed his most recent contract, he was the most valuable possible trade commodity in the game. Now it's less so, and NL teams should beware. I'm not suggesting that they trade him just to get him out of town or just to get Manny some DH at-bats, but if I read the perception around the game well, the Sox could solve many of their roster construction problems in a single move that way.

This assumes that you want to use the DH slot as most teams do, which is as a way to rest a player but still keep his bat in the game. I think a good DH one of the last offensive ineffeciencies that can be exploited, as most teams do not get very good production out of their DHs.

If we accpet that most players suffer a pinch hitting or DHing penalty at the plate when they're not playing the field, then it makes sense to treat a DH who does not suffer that penalty (or actually finds it to be an advantage) as playing a legitimate position. Over his career with the Sox, Ortiz hits much better when playing as the DH, so to does Hafner, and the latest incarnation of the Big Hurt (who admittedly would probably find it hard to work one half of the wheel chair with a 1Bman's mitt on). More teams trying to leverage this advantage: Toronto (by aquiring Thomas), as well as Detroit (Sheff), and Oakland (using Piazza to make up for loosing Thomas).

OTOH, forcing players who suffer the DH penalty into the slot is harder to do. Manny suffers a decline in his OPS when DHing; Giambi has a similar problem, which is why the Yanks have tried to keep him in the field for as long as possible. The effect on overall offense would seem to justify this; Manny in LF can still play there, but the concern with Tiz/Pronk is injuries that remove their bats completely.

If we accpet that most players suffer a pinch hitting or DHing penalty at the plate when they're not playing the field, then it makes sense to treat a DH who does not suffer that penalty ... OTOH, forcing players who suffer the DH penalty into the slot is harder to do. Manny suffers a decline in his OPS when DHing; Giambi has a similar problem, which is why the Yanks have tried to keep him in the field for as long as possible.

The problem with quantifying the "DH penalty" is that a player who appears both at DH and at a regular position during the year is more likely to be DH if they are not at 100%. That is, it is often used to get a bat into a game when they are slightly injured but not so bad as to require the DL or a full day off. Perhaps there is such a penalty for a player who may not be as focused on the game or perhaps overly focused on each at bat, if they only DH, but the data is too polluted to show it.

you're using chances as defined by STATS ZR, which ignores many line drives. Under a more liberal definition of all the air balls hit at least 200 feet, you'd get more like 500 chances a year. (HBT, now using the new "Dewan" ZR, shows 494 chances per AL team in 2006 because they count all those line drives with a "low" probability of being caught.)

I am. I think that's correct. I think having a baseline of 50% chance is a reasonable line as long as you subsequently reward for plays OOZ (however defined).

There's an interesting data set available here to illuminate this point.

I am unsuccessful in accessing that paper, so the direction isn't clear to me. Any input on that?

But as Darren noted in #27, it likely overcorrects

It may overcorrect. And the overcorrection is going to be very very small. Maybe 3 runs (doing some Q&D math - 40 plays 100% conversion vs 80% conversion). Plus more than a few of those balls would actually not be BIP - they would be home runs.

Thanks for the input.

1.865 catches per game in LF for the average LF'er (in 2006). I define a "chance" for a fielder as the number of balls turned into outs that an average fielder would make given the same distribution of balls in play. You (anyone) can define a chance any way you want as long as it is clear what you are using it for and the definition fits the use. When you are using a ZR system, you can define a chance as a "ball in zone." When you are using a system that looks at all balls regardless of where they are hit, there is no logical way to define a chance - technically, every ball is a "chance," especially of the league average catch rate for that ball is greater than zero. You are left with arbitrarily defining "chances" and I choose to use the definition I do. I suppose that a true "chance" would be somewhat higher than the number of balls that an average fielder catches. Maybe the number of balls that the best fielder in the world would catch? Who cares? It doesn't matter. Regardless of what definition you use, as long as you compare a player's "chances" to a league average fielder you can figure out "defensive games" if that is what you are trying to do. And as long as you approximate how many balls a player has a reasonable chance of catching, you can use that to figure standard deviations, standard errors, etc., for a player's UZR (or whatever metric you are computing).

Fair enough. 1.865/0.87 = 2.14

How many PAs in each case, Kevin?

The problem with quantifying the "DH penalty" is that a player who appears both at DH and at a regular position during the year is more likely to be DH if they are not at 100%. That is, it is often used to get a bat into a game when they are slightly injured but not so bad as to require the DL or a full day off. Perhaps there is such a penalty for a player who may not be as focused on the game or perhaps overly focused on each at bat, if they only DH, but the data is too polluted to show it.

That goes to part of my point, though, that the DH spot tends to be under-utilized relative to it's potential to help an offense. It would take a lot more study to find completely "clean" stats on the matter, but it's also a case of which is more valuable, having a DH like Papi or the roster flexibility to get an injured guy a few more ABs (instead of letting him heal) and giving AB's to Phil Nevin in the meantime?

I don't know. The difference looks pretty small to me.

I was going off of memory; the last time I looked at the 3 year splits they were '03-'05. IIRC the difference was a bit more pronounced (although I don't exactly think that 40 points of OPS is anything to sneeze at), with Manny loosing about 100 points of OPS as a DH and Ortiz loosing another 100 OPS as a 1Bman. When adding in '06 and removing '03 (nevermind correcting for my possibly faulty memeory) Ortiz has an identical 1.012 OPS as a DH and as a 1Bman (in a much smaller sample size).

To a certain extent we're never going to get an ideal sample because of the Sox usage patterns, but over their careers here are their numbers:

Papi as 1B: .282/.364/.506 (.870) (935 PA)
Papi as DH: .286/.378/.568 (.946) (3270 PA)

Manny as LF: .311/.411/.619 (1.030) (2820 PA)
Manny as DH: .315/.411/.579 (.990) (1074 PA) (leaving out Manny as a RF completely.)

Hypothesis: players who don't normally DH tend to hit worse when they do, as they only DH when injured or at less than 100% health.

A 95 at-bat sample spread over several years is essentially useless though.

Hypothesis: players who don't normally DH tend to hit worse when they do, as they only DH when injured or at less than 100% health.

As I am sure you know, that's been the explanation for at least since the internets have been in use.

Yes Dial, I know. I just don't know if the hypothesis has ever been tested in any way.

Yes, when we mentioned the "DH penalty" in the book, we posited the theory that it may be because players are often DH'd when too injured to play the field. There is no doubt that among players who DH and play the field, they are less healthy when they DH. IOW, if there were no "inherent" (not having anything to do with health) DH penalty, we would expect players to perform, on the average, worse when DH'ing. However, we still don't know if there is an "additional" DH penalty having to do with sitting on the bench, not being "loose" or "focued", etc. OTOH, for all we know, not playing the field may increase a player's hitting performance. It would be hard to get an answer, even if someone tried to use DL data to get a sample of players (who DH and play the field) who did not appear to be injured at all. Maybe even healthy players who are "tired" DH in order to get a day or two off from the field. One would have to adjust for age as well (I don't remember off the top of my head if Tom adjusted for age in the study reported in The Book, or if he always used same year data, as players are more likely to DH as they get older.

To make matters worse, to the extent that it is a mental thing, the DH penalty may only affect certain batters, or may affect certain batters more than others.

To summarize, here's what we have from each for 06 for runs cost per full season of play:

Arthur: -12 (using an average of his and AROM's park adjustment)
Dial: -12
Chone: -35
MGL: -32

I was the one who averaged my adjustment (14 runs) with Joe's. That -35 runs is completely unadjusted. My projection for Manny is -14 runs.

AROM,

That -35 (should read -31) comes from this ST thread. In it, you wrote:

I have Manny, unadjusted, at -31 runs with 248 chances.

So using my adjustment by this article, he's -22...

To which I replied:

To be clear, this is for the 248 chances, right? If we were to prorate to 150 games, the numbers would come out to about -31...

Which you agreed with. So your numbers, using your adjustments got him to -31 per 400 chances (your number as well).

These numbers were in reference to Manny's 2006, with your park adjustments (not Arthur's or a combination of the two). They were not a projection, which is what you supplied in the post above.

Darren,

I had to track down my latest defensive numbers through a few hard drives and flash drives. Looks like I never got around to posting them last year.

Anyway, I have Manny, in his actual playing time, at:

2006 -16
2005 -14
2004 -6
2003 -2

With a -13 projection. I don't know why I had -14 before, maybe a formula tinker and rounding difference. What this tells me is someday I need to learn some organization skills.