Categories

Archives

Value of Brendan Harris’ Defense

ubelmann November 13, 2009, at 12:00 pm 10 LTEs - (Sorry, but the time to add LTEs has expired) , ,

In what is apparently becoming an ongoing series, this is in the spirit of similar posts dealing with Denard Span and Nick Punto, although this time we run across someone who is a poor defender.

Even though Harris has been kicking around for a while, because he's been a part-time player, we don't even quite have three full seasons of data on him. I'm going to ignore his 5 games at first base as too small to tell anything, but at 2B/3B/SS, he has 356 total defensive games. Those are the same positions that Punto covers, so essentially everything I said there applies here.

Let's take a look at what we have to work with:

-6.8 UZR/150G -- 2B
-19.6 UZR/150G -- 3B
-9.9 UZR/150G -- SS

This is a bit different than Punto's data in that we could imagine Dirty Harry's got a poor first step that hurts him more at 3B and can run down stuff in the middle infield to make up for a poor first step. And he's worse at SS than 2B, which also seemingly makes sense.

But Harris largely came up as a third baseman, and I'm unconvinced that we have enough data here to really say that he's particularly bad at 3B compared to 2B. So we go with the standard assumptions here that he's equally good at 2B/3B and 5 runs/season worse at SS than 2B/3B, and we'll normalize to 2B/3B.

-6.8 UZR/150G, -4.1 UZR, 90 G -- 2B
-19.6 UZR/150G, -9.2 UZR, 74 G -- 3B (already normalized to 2B/3B)
-4.9 UZR/150G, -6.3 UZR, 192 G -- SS normalized to 2B/3B
-------------------------------------------------------------------------
-8.2 UZR/150G, -19.6 UZR, 356 G -- All IF positions normalized to 2B/3B

Then we need to regress this to the mean, which at first seemed kind of generous to me, but I guess enough people have given Harris a shot to play defensively challenging positions that it's not out of the question to regress him to an average 2B/3B. Assuming 94 additional games where he plays league average defense at 2B/3B we get:

-8.2 UZRR/150G, -19.6 UZR, 356 G -- All IF positions normalized to 2B/3B
0.0 UZR/150G, +0.0 UZR, 94 G
-------------------------------------------------------------------------
-6.5 UZR/150 G, -19.6 UZR, 450 G -- regressed performance in IF normalized to 2B/3B

That essentially makes Harris 6-7 runs/season below average at 2B or 3B and 11-12 runs/season below average at SS. That's better than I thought he'd turn out when I started this, but still enough worse than Punto that I'd generally start Punto over Harris. (And I still think that Harris could go all Casey Blake on the league if someone let him settle in at third base, but I don't know if I want the Twins to be the team to make that move.)

Now that I've done three of these, I suppose I can start a little ranking list. In addition to the multi-position players I've done so far, I'll include the career defensive rating for single-position players who have more than 450 games at their position. (So no adjustments either way are necessary. Except an aging curve, I suppose, but I don't have a consistent method to do that at the moment.)

Twins defensive rankings so far (position-neutral):
+18.7 R/150G -- J.J. Hardy (+11.2 UZR/150G at SS)
+17.8 R/150G -- Nick Punto
+12.5 R/150G -- Joe Mauer (??? relative to average catcher)
-2.3 R/150G -- Denard Span
-4.0 R/150G -- Brendan Harris
-9.1 R/150G -- Justin Morneau (+3.4 UZR/150G at 1B)
-16.0 R/150G -- Delmon Young (mostly covered here, except regression)

Multi-position players left on the docket: Cuddyer (IF and OF separately, so that'll take a bit more time), Tolbert (probably mostly regression on his SSS), and Kubel (with thoughts on his time at DH.)

Value of Nick Punto's Defense

ubelmann November 12, 2009, at 12:00 pm 11 LTEs - (Sorry, but the time to add LTEs has expired) , ,

In the same spirit as my post on Denard Span's defense, let's take a look at Nick Punto's defense. He also has performances at different positions that don't make a whole lot of sense. To wit:

+3.9 UZR/150G -- 2B
+19.9 UZR/150G -- 3B
+18.1 UZR/150G -- SS

Typically 2B and 3B who switch off between positions play them relatively evenly and 2B/3B/SS who switch off between positions play SS about 5 runs per season worse than the other positions. But here we see Punto scoring much worse at 2B than 3B, even though they should be relatively equally demanding, and just as good at SS as 3B, which doesn't make a whole lot of sense either.

Punto is in an interesting situation, similar but different from the one that Span is in. Punto has well over 3 full seasons of UZR data to draw from (594 defensive games is roughly 4 full seasons), but he doesn't have close to 3 full seasons at any one position (208 games at 2B is well short of the 450 games in 3 full seasons.)

I do believe that infield is a materially different situation than the outfield in that all outfield positions require essentially the same set of skills in more or less the same ratios, whereas different infield positions require infielding skills in different ratios. However, just using the data, it seems to me that you'd need at least three seasons of data at each position to really establish any kind of certainty that a player is, say, an elite shortstop but only a so-so second baseman. And thinking about Punto's tools, it doesn't make a whole lot of sense to me that he'd be rated so much higher at SS than at 2B, or so much higher at 3B than at SS. He's got a strong arm and pretty sure hands. He is also a pretty decent runner, though I don't see him as having necessarily the best first step in the world. Which wouldn't necessarily make him a great third baseman, but should make him a pretty good middle infielder, though I can see no reason that he should be better at SS than 2B. (And even if you see his tools differently, I can't imagine a combination that makes him good at SS and 3B but not at 2B.)

So with all that said, I'm going to assume that Punto is in fact equally good (compared to league average) at 2B and 3B, and 5 runs/season worse at SS than at 2B/3B. Under these assumptions, we can normalize all of his fielding data against a league-average 2B (since I see that as his most likely position next year) and see what it all adds up to.

+3.9 UZR/150G, 5.2 UZR, 208 G -- 2B
+19.9 UZR/150G, 25.4 UZR, 192 G -- 3B normalized to 2B
+23.1 UZR/150G, 29.9 UZR, 194 G -- SS normalized to 2B
------------------------------------------------------------------------
+15.3 UZR/150G, 60.5 UZR, 594 G -- All IF positions normalized to 2B

I won't do a regression to the mean here in the way that I did with Span because the data all combined together is more than three season's worth of data.

That makes Punto roughly a +15 runs/season fielder at 2B or 3B, and a +10 runs/season fielder at SS, at least if we're using UZR as our inputs. I believe that other systems don't rate him as highly, but this data doesn't strike me as being outrageous. If you thought he was more like +10 at 2B/3B and +5 at SS, I wouldn't blame you, but I think it's fairly clear that he's an above-average up-the-middle defender, which definitely has value. (And if you find him difficult to stomach because of his headfirst slides into first base and boneheaded bunt attempts, I also wouldn't blame you.)

Value of Denard Span's Defense

ubelmann November 8, 2009, at 12:00 pm 19 LTEs - (Sorry, but the time to add LTEs has expired) , ,

Continuing with the aftermath of the Gomez/Hardy trade, let's take a closer look at the value of Denard Span's defense. He's impressed me in the field, especially in the corners, but looks can be deceiving. On the other hand, small sample size numbers can be deceiving, and he doesn't have a full season's worth of data at any one outfield position. Span's really made such an impact over the last couple of seasons that it's difficult to remember that he doesn't even have two full seasons of experience at all OF positions combined.

UZR is about as good as any other defensive metric and it's easily available. It doesn't necessarily tell a very consistent* story, though. Let's take a look at his numbers for the last two years combined:

-13.8 UZR/150G, -7.7 UZR, 83 G -- CF
+16.7 UZR/150G, +6.0 UZR, 55 G -- LF
+5.8 UZR/150G, +4.8 UZR, 117 G -- RF

*And I'm not referring to how they calculate their ratios, though I could be. There seems to be some round-off issue that I don't understand, but it's not that big of a difference since we are looking at the data to get a good idea of how he is on defense, not to measure his defensive contributions to a millionth of a run.

He grades out better in the corners than in CF, which makes sense. But it doesn't necessarily make any sense that he should be so much better in LF than in RF, and furthermore, the typical difference in value between corner OF and CF is about 10 runs per 150G, and the differences here are much larger than that.

So what I propose is to suppose that Span is 10 runs worse in CF than in the corners (and that he is equally good in each corner) and combine the data under those assumptions. This basically gives us:

-13.8 UZR/150 G, 83 G -- CF
+6.7 UZR/150G, 55 G -- LF data normalized to CF
-4.2 UZR/150G, 117 G -- RF data normalized to CF
---------------------------------------------------------------
-8.5 UZR/150G, 255 G -- All data normalized to CF

Then, because it's still a small sample size, we should probably figure out a way to regress the data towards the mean. To do that we have to figure out what mean we are regressing towards. I would say that it's reasonable to regress Span's data to that of an average-fielding CF. Scouting-wise, no one seems to think he looks awful in CF (except perhaps in comparison to Gomez, who is well above average), and that seems like reason enough to regress him towards an average center fielder.

Usually the number that UZR proponents throw around is that you need three seasons of UZR data to make a good prediction for the next year's performance. I'll take 3 years to mean 450 games. So all we'll do here to do the regression is assume that in addition to the 255 games that he did play, Span played 195 games (the difference between 450 and 255) at exactly league average. That gives us:

-8.5 UZR/150G, 255 G, -14.5 UZR
+0.0 UZR/150G, 195 G, 0.0 UZR
------------------------------------------
-4.8 UZR/150G, 450 G, -14.5 UZR

Overall, then, I guess I see Span as about 5 runs below an average center fielder over the course of a full season and 5 runs above an average corner outfielder over the course of a full season. He could prove me wrong (or he could improve or regress) next year, but I think that's a good estimate of where he's at right now.

The Fielding Bible II

ubelmann March 11, 2009, at 2:52 am 17 LTEs - (Sorry, but the time to add LTEs has expired) , , , ,

I guess somehow I missed this until, but John Dewan (the plus/minus guy) has released the second edition of his Fielding Bible. The occasionally stubborn J.C. Bradbury has a tiny rundown at his blog, including a link to an excerpt of FBII.

I'm in the process of reading the excerpt now, with plans on buying the book if I can find a way to justify expanding my baseball bookshelf. An initial concern:

Click here to continue reading The Fielding Bible II...

DeRosa To Indians/Aaron Miles to Cubs/Improving The 2009 Twins

ubelmann December 31, 2008, at 4:57 pm 46 LTEs - (Sorry, but the time to add LTEs has expired) , , , , , , , ,

per ESPN:

DeRosa has been traded to the Cleveland Indians for minor league pitchers Jeff Stevens, Chris Archer and John Gaub. DeRosa will likely slide into Cleveland's opening at third base.

I'm sure this will irk fans who long for the Twins to do something and especially those who would like the Twins to add someone they've heard of before, but DeRosa wouldn't really be a significant upgrade.

In related infielding news, the Cubs signed Aaron Miles. Elsewhere, I've seen the total value of the contract listed at 2 years/$4.9M.

For these two infielders, some internal options, and other names we've heard this offseason, let's look at projected wOBA (as determined by Marcel the Monkey), and career OPS+.

.394 wOBA, 147 OPS+ -- A-Rod
.350 wOBA, 97 OPS+ -- DeRosa
.335 wOBA, 105 OPS+ -- Blake
.329 wOBA, 97 OPS+ -- Harris
.323 wOBA, 93 OPS+ -- Buscher
.308 wOBA, 77 OPS+ -- Miles
.299 wOBA, 74 OPS+ -- Punto
.290 WOBA, 67 OPS+ -- Cesar Izturis

DeRosa's a little younger than Blake, which is likely the main reason his projected wOBA is better than Blake's. DeRosa has been much better than his career averages over the last three years, which is likely the main reason that his projected wOBA is better than Harris'.

First we see that none of the non-Rods are elite hitters and Punto isn't so bad when you compare him to other guys who have gotten similar 2-year deals this offseason. If we're considering these guys as potential full-time players, let's look at how many runs above average their wOBA would make them over 550 PA:

29.7 -- A-Rod
8.6 -- DeRosa
1.4 -- Blake
-1.4 -- Harris
-4.3 -- Buscher
-11.5 -- Miles
-15.8 -- Punto
-20.1 -- Izturis

DeRosa is better at the plate than Harris and Buscher, by maybe about one win over the course of the season. Then it becomes a matter of how well everyone can field. Defensive stats are a bit of a mess for a number of reasons, but let's take a look at career UZR per 150 games at third base for these guys. There are definitely sample size issues here, and I don't think all of these ratings make sense, but it will give us a place to start the conversation.

18.7 -- Punto
8.2 -- Izturis
0.9 -- A-Rod
??? -- Miles
-3.4 -- Blake
-4.1 -- DeRosa
-12.8 -- Buscher
-18.3 -- Harris

First things first--I think that Punto's rating and Harris' rating are each too extreme. No one else seems to rate Punto that high at 3B and I would guess that he's closer to a +10 defender at 3B. Harris, on the other hand, has started only 39 career games at 3B, so there are extreme sample size issues with that number. UZR has him as a -3.9 runs/150G defender at 2B and a -9.9 runs/150G defender at 3B. Compared to other metrics, I think that's generous, but given his tools and the information from 200 games started at 2B/SS, I think it's reasonable to expect that Harris can be a -5 defender at 3B. Optimistically, I would guess he's average and pessimistically, I think -10 is about right at 3B.

Taking those things into consideration, and rounding to emphasize that these are rough measures, here is how I would rate those players as defenders at third base, in runs above average:

10 -- Izturis
10 -- Punto
0 -- A-Rod
0 -- Miles
-5 -- Blake
-5 -- DeRosa
-5 -- Harris
-10 -- Buscher

Using those defensive estimates, we get, for overall value at third base, in runs above average:

30 -- A-Rod
4 -- DeRosa
-4 -- Blake
-6 -- Punto
-6 -- Harris
-10 -- Izturis
-12 -- Miles
-14 -- Buscher

Of the group, I also have the least confidence in Buscher since he had an awfully weak history before the last couple of years and I'm most concerned about his defense, which could very well be worse than -10.

The more I look at it, the more I would lean towards making Harris the full-time 3B, with Buscher playing against tough right-handed pitchers or when we're not as concerned about infield defense. Harris seems more or less to be Casey Blake from 6-8 years ago, and we doesn't cost us anything to acquire him.

I think DeRosa would have been an upgrade. But he would have been a small upgrade, and for an extra $5.5M in salary, plus giving up more talent than the Indians gave up. I'm not sure at all about how good those minor league pitchers are (I suspect they aren't very good), but I'm sure that they have some value. $5.5M and low-value prospects are more than I would have given up for one extra win.

The Twins did a lot last offseason, and they did have done very little this offseason, but given their performances in 2007 and 2008, I think that makes sense. In 2007 we had a number of below-replacement-level performances. Upgrading from awful to mediocre is the easiest kind of improvement that you can make.

In 2008, a big reason that the Twins improved is that they had barely any below-replacement-level performances. Monroe and Everett, for instance, were annoying, but as a group they were only as far below replacement level as Darnell McDonald, Matt LeCroy, and Chris Heintz were (as a group) for the Twins in 2007. Mike Lamb was all kinds of fail this year (especially when you include his glove), but that's nothing compared to the devastation that Rabe, Garrett Jones, Luis Rodriguez, Rondell White, Alexi Casilla, and Nick Punto brought us the year before.

All totalled, the Twins improved from -76 runs in below-replacement-level performance in 2007 to just -19 runs in below-replacement-level performance in 2008. That's a 5-6-win improvement in just replacing your worst players with barely adequate players. That's very nearly as good as the difference between A-Rod and a replacement-level third baseman.

I think that's a pretty overlooked part of the '07 to '08 improvement, but it also means that we've more or less exhausted the cheap and easy ways to improve the team--at least in position players. We ought to be scouring the waiver wires and looking for interesting NRIs to shore up the bullpen and rotation depth. Even there, I don't know how much you can realistically expect improvement. On results, Bonser sucked last year, but I suspect he'll be at or above replacement level this year. Bass was also bad, but he wasn't even below replacement level based on runs allowed--so I wouldn't expect for miracles in improving the bullpen.

The Twins have a roster that can be substantially improved for 2009 in three main ways:

1) Fleece someone in a trade
2) Trade future value for present value
3) Spend $$$$ to improve spots where we are okay but not great

The first option always sounds great, but you can't always depend on deals with suckers to come through. The second option probably doesn't make sense for the Twins because money is a very scarce resource for them and prospects usually costs less money than players you would acquire in a trade. The third option doesn't work because we're not lucky enough to be sitting on a 25-million-resident gold mine.

I didn't really plan on coming back to this, but here I am again. For as much as some people want to be delusional about the advantage that the Yankees have, it's ridiculous to think that small-revenue teams like the Twins aren't at a significant disadvantage compared to large-revenue teams when it comes to assembling their roster. I can hope that the Twins can turn fool's gold into gold, but it's not reasonable to expect that, and I see no obvious, significant upgrades for this team.

Terry Ryan built the basis for this team by being patient and not trading the future for one wild run at a ring. Bill Smith would be wise to follow that precedent. Our offseasons are bound to be boring not because of our management, but because of our budget. Until MLB decides to do something about that, it's just the way it is.