Thursday, March 10, 2005

New Home

And so it was written that March 10, 2005 shall henceforth be forever known as the Day of the Great Baseball Blog Exodus.

I'm out of here; the Fourth Outfielder has moved to All-Baseball.com. Please check me out at

http://www.all-baseball.com/fourthof/

You may change your bookmarks if you have any.

Wednesday, March 09, 2005

Calls to the Bullpen


Tuesday, March 08, 2005

Notes

Two Dodger-related items at Baseball Prospectus (both subscription only) today. The first is Will Carrol's team health report. Will does a great job on these, but to be frank I didn't really learn anything new from his player comments. His discussion of Frank Jobe's rotator cuff work was pretty interesting, though. I think Will misstated his line about Greg Miller - he was out last season and will miss several months now, but last I heard he's expected to pitch this year. Also, I don't think Will is familiar with the work Wilson Alvarez has done with the Dodgers.

And the red light for OP and green light for Lowe - though both things you and I already knew about - caused me to briefly deliberate once again on how baffling it is that the near universal sentiment is that Perez at $8m per was a good and necessary deal while Lowe at $9m is terrible. Given the choice, I'd probably take the Perez deal, but it's no slam dunk. They're both overpaid.

The second item is Joe Sheehan's mention of Dioner Navarro as one of his "Not My Guys." Sheehan rightly states that Navarro's offense is tied closely to his batting average. However, his mentioning Navarro's .263 average across two levels comes across as extraordinarily naive. A 20-year-old putting up an above average BA and OBP in Double-A is pretty good, and the .250 he hit in AAA is, given his age, not shabby. The idea that Navarro is the next catching superstar is silly, but that doesn't mean he can't or won't be a very good player. The Ivan Rodriguez comparisons I've seen were, in my mind, mostly about his defense. I wouldn't expect Navarro to put up a career .184 ISO as Rodriguez has, but I also would expect him to walk more frequently than Rodriguez and he's got a good shot of hitting for average nearly as well as Rodriguez. And while it's unlikely his defense will be Rodriguez caliber, he could still be pretty good with the mitt.

Also, Sheehan's invocation of Mike LaValliere is puzzling; I don't know about you, but I'd take LaValliere's first six seasons in a heartbeat. He was an above average hitter at a defensive position and most metrics I've seen have shown his defense to have been excellent. Is Navarro less of a prospect because LaValliere was never allowed to be a full time player? Even so, the comparison is pretty poor in my eyes. LaValliere was a left-handed hitter who didn't sniff the majors until age 23 and didn't see any non-September playing time until age 25. Navarro has been an above average AA hitter at age 19 and 20. Not adding up in my mind.

Monday, March 07, 2005

Can Choi Keep Up his .819 OBP against Southpaws?

Before I continue, a typo that had me laughing too hard to ignore it... From the latest Gurnick mailbag:

According to general manager Paul DePodesta, if Choi winds up with statistics similar to his season-ending totals of last year -- .251 batting average, .819 on-base percentage, 15 home runs and 46 RBIs -- that would be sufficient.

Did he really have a .819 on-base percentage? That would suffice in my book. A .251/.819/.251 season would be MVP caliber.

In any event, I wanted to take a look at a question which has come to th fore a bit more recently. Should Hee Seop Choi start on days when the Dodgers face a southpaw?

We can assume that Jose Valentin is not going to start on days when a left-handed pitcher is throwing, so that means that one of the APerez/Nakamura/Saenz trio would be starting anyway. That means the potential platoon mate would be the second-best of that group.

To get a quick and dirty idea of what kind of impact we're talking about here, I'm going to use the average projection from PECOTA, ZiPS, and Marcel that I found a few weeks back. In the quick and dirty spirit, I'll use each player's Gross Production Average. Choi's is .281, Saenz' is .265, Perez' is .248, and Nakamura's is .250.

To project each player's performance against southpaws, we can use the universal platoon ratio, which, given the sample size for each individual player, is normally a better predictor of future platoon ratio than past platoon ratio. Left-handed batters hit about 1.17 times better against right-handed pitchers and right-handed batters hit about 1.09 times better against left-handed pitchers. We can reverse engineer our general projections based on how many plate appearances against each type of pitcher they had in the past, since each player's projection is based on seasons where they had a certain number of PA against LHP. I don't have the actual numbers for Nakamura or Perez, but I'll assume that since they've been everyday players they've had the normal ratio of about 28% PA against LHP, although ideally I would find the actual data for Japan and AAA... Anyway, here's the table, with plate appearances being from 2002-2004:

RHP PA LHP PA %LHP ProjGPA PlatoonRatio ProjLHP ProjRHP
Choi 647 71 10 .281 0.85 .242 .285
Saenz 160 146 48 .265 1.09 .277 .254
Perez 28 .248 1.09 .264 .242
Nakamura 28 .250 1.09 .266 .244
Using this method, Saenz comes out on top, followed by a virtual tie between Perez and Nakamura. However, Saenz is the one least likely to play third base, so we can figure that it will be one of the others playing in Valentin's stead. So let's compare Saenz to Choi. Saenz' has a .035 edge in GPA using these numbers. A rough run estimator for GPA is Runs = (GPA_1 - GPA_2)*1.2*(plate appearances). 698 starts were made in the NL last season, or 27% of the 2590 games played. The average starter lasts 27 batters, so we're talking about three plate appearances a game in 27% of the Dodgers 162 games. 3 PA times 44 games equals 132 plate appearances at stake. So 132 PA * 1.2 * .035 GPA difference equals five and a half runs, or half of one win.

So, the difference between Choi and Saenz starting, offensively, is half a win. I don't think there's too big a spread in their defensive talent, so let's assume it's simply a question of offense.

Now, there's another question at stake here: will playing regularly against left-handed pitching improve Choi's long-term ability, both in general and against left-handed pitching? The answer is almost certainly yes, but the magnitude of the impact is a different question. On the one hand, it's not a big deal to carry around a right-handed backup and platoon partner at first base. That's perhaps the most easily replaceable role in baseball. On the other hand, having to often remove your first baseman in the sixth or seventh inning for a pinch-hitter creates a number of strategic limitations, so it's certainly preferable to have your premier left-handed hitters have better than average platoon ratios. And while keeping a decent right-handed first baseman on the roster is easy and fairly inevitable, it's certainly better to not have to depend on one and to not have to spend a little extra for a very good one. And some of the value added from Choi getting experience against southpaws will be this season when he has to face left-handed relievers and it's not strategic or possible to remove Choi from the game.

I would argue that Choi should probably be the default starters against left-handed pitchers but that Saenz should get the nod in high-leverage games. That is, when the Dodgers are playing a clear rival for a playoff spot or are in a tight spot late in the season - in other words, when wins are at a premium - Saenz should face the southpaw.

With regards to pinch-hitting, the same philosophy should apply. Eighth or ninth inning of a close game and a lefty's on the mound? Bring in Saenz. Three-run lead in the seventh? Let Choi hit. Also, the possibility of the next time through the order has to be factored in. Saenz has a slight edge over Choi if the task is to one PA against a southpaw and one against a righty. However, if there's reason to believe that there will be another trip through the order and that Choi's spot in the lineup will occasion a higher leverage plate appearance the next time, then it makes more sense to leave Choi in the game.

Sunday, March 06, 2005

(Bad Risky Business/Joel Goodson Pun Omitted)

Joel Guzman has been generated an awful lot of Miguel Cabrera comparisons. Are these comparisons accurate?

Cabrera's development saw a rapid spike in 2003. Before then, he was clearly talented and young enough to develop, but he hadn't put up impressive numbers. Playing as a teenager in rookie league, low A, and high A, Cabrera held his own in each league but never excelled. At each level he put up offense just above league average. Every component of his performance in rookie ball at age 17 was just about average. His age 18 season in low-A saw his K rate become substantially better than average, but that was offset by smaller declines in his walk and hit rates. Promoted to high-A at age 19, he kept his solid K rate but saw his BB rate sink a bit further back. This was more than offset by a sharp increase in doubles power.

At age 20, starting the season in AA, Cabrera broke out. The doubles power he'd shown the previous year remained, but it was augmented by a huge improvement in his home run power. On top of that, Cabrera maintained his good K rate while putting up an excellent walk rate. He also saw a sharp increase in his singles rate. Cabrera's dominance of the Southern League prompted the Marlins to skip him past triple-A, and you know the rest of the story from here.

Guzman debuted in 2002 at age 17. Playing briefly in the Pioneer League and then even more briefly in the GCL, Guzman held his own against older competition but put up subpar numbers overall. He demonstrated patience at the plate, putting up an above average walk rate and a terrible strikeout rate. He hit the ball very hard, generating very good power numbers per balls in play. However, the K rate was way too high for Guzman to be productive.

At age 18, Guzman debuted in Sally ball at South Georgia. Once again, Guzman's power relative to the league was outstanding. It was tough to notice, though, in part because the Sally League is tough on hitters and in part because his K rate - while improved since rookie ball - was still very bad. On top of that, Guzman's hit rate on balls in play lagged below average and his walk rate was microscopic. Promoted to high-A Vero Beach at mid-season, Guzman's power dropped off but was still well above league average. His singles rate, K rate, and BB rate all improved slightly after the promotion.

At age 19, Guzman saw his breakout. Starting the season at Vero Beach, Guzman put up tremendous power numbers and a much improved BABIP. His K and BB rates again showed incremental improvement, but both remained well below average. After being promoted in July to Double-A, Guzman's power numbers relative to the league fell off a bit but remained outstanding. Meanwhile, his K and BB rates both improved.

In all, I don't think the Cabrera comparison is very apt. Cabrera's issue wasn't really conquering the strike zone, as he managed to do a fine if subpar job of that at each level. Guzman, on the other hand, has neither drawn walks consistently nor avoided striking out, although he's shown significant improvement in that area the past two seasons. Unlike Guzman, Cabrera did not consistently display solid power as a teenager and his challenge was adding home runs and doubles to his attack.

To give you an idea of what I mean, here's a comparison of a few metrics that I think tell this story well. I might get around to introducing these numbers and their respective scales in a little more detail later, so I'll give you the short version here. Each one is like "OPS+" in that it's expressed as a percentage of the league average. lwr+ measures a player's overall offensive value using a formula based on Tangotiger's linear weights ratio. wx+ measures the value of a player's extra base hits per batted ball. b+ measures the value of a player's batted balls. t+ measures the value of a player's plate appearances that don't result in a batted ball - HBP, BB, and K.


Joel Guzman Miguel Cabrera
Age League lwr+ wx+ b+ t+ Age League lwr+ wx+ b+ t+
17 Pioneer 95 137 126 60 17 GCL 103 96 105 91
17 GCL 87 87 84 102
18 SAL 92 156 120 27 18 Midwest 103 96 102 94
18 FSL 92 130 115 33
19 FSL 134 200 163 55 19 FSL 109 131 111 96
19 Southern 117 164 132 66
20 Southern 172 197 182 129


To me, Cabrera just is not a very good comparison.

The issue for Guzman is not whether he can cut down on his strikeouts but rather how he'll cut down on his strikeouts. If he can start hitting more pitches but continue to hit them hard, he'll have no problem - in other words, if he can keep mashing while becoming less selective, there's no question that he's an outstanding prospect. Conversely, if his current selectivity is necessary for his power, he'll need to develop a better feel for the strike zone and turn a lot of those strikeouts into walks. So far, he's mainly done the latter, as his batted balls per plate appearance hasn't fluctuated much but his K and BB rates have improved. If he continues to do that - and continues to smack the balls he does hit - he'll be just fine. If Guzman can't cut down on his strikeouts without compromising his ability to pound the ball, then his future is in jeopardy.

Friday, March 04, 2005

Franquelis Osoria

Most ardent Dodger fans already know about Ryan Ketchner, a swingman in the Seattle organization acquired last season for Jolbert Cabrera. Ketchner looks like he's not too far from being a solid third or fourth starter if his recovery from ulnar nerve transposition goes well.

Franquelis Osoria arguably had a better season than Ketchner last season, but he's suffering from IES (inflated ERA syndrome). He put up a 3.67 ERA in 81 IP at Jacksonville and a 6.48 ERA in 8 IP at Las Vegas. Those numbers don't make him look like a top prospect, considering that at 22 he wasn't particularly young for the competition. However, a look at Osoria's peripherals shows much more promise, as in 2004 he had an excellent ratio of 76 K to 19 BB while only giving up two home runs in 89 innings. By contrast, Ketchner pitched 126 innings with 102 K, 36 BB, and 11 HR.

Using their weighted three year numbers, Ketchner and Osoria are fairly similar; they have nearly identical walk/HBP rates and hit rates on balls in play. Ketchner is the better strikeout pitcher, though they're both good in that department. Osoria has been much better at keeping the ball in the park. Osoria is half a year older. There's a good reason Osoria was added to the 40-man roster last fall to avoid the Rule 5 draft.

Osoria doesn't project as a world-beater, but he's got a decent shot to be a solid contributor at the major league level in the next few years.

Thursday, March 03, 2005

Memory/Lack Thereof Lane...

How did I not remember this?

Florida Marlins traded Craig Counsell to Los Angeles Dodgers in exchange for a player to be named later (June 15, 1999); Florida Marlins received Ryan Moskau (July 15, 1999). (from retrosheet)

Craig Counsell played for the Dodgers? The Davey Johnson Dodgers? Counsell was terrible, of course (he is Craig Counsell); a no-slug, all-OBP guy whose walk rate isn't especially high and who depends on a lot of groundball singles isn't a good match for anybody, and Dodger Stadium in particular is a poor match. He was released March 15, 2000. The weird part is, this is about as close to a "win" as a one-for-one player swap to get Craig Counsell can be: Moskau, a first baseman whom the Dodgers picked in the 6th round the previous June, didn't amount to anything. Still, even giving up nothing to pay Counsell's salary isn't worth it.

The Dodgers ended up dumping a worthless utility infielder of their own shortly after releasing Counsell. On April 1, they dished Juan Castro to the Reds for Kenny Lutz. Lutz was drafted by the Reds out of Cincinnati in the fourth-round the previous June, and he never had any professional success. Fortunately, the Dodgers also picked up $150,000 in the trade, making it an unequivocally good deal for the blue crew; Castro was about as useless as possible for a player on a major league roster from 2000 to 2002, before slightly improving from terrble to bad the past two seasons. Now the Twins are giving him money to use up outs.

But the Dodgers weren't done excising their bizarre surplus of sub-mediocre middle infielders; they dealt Jose Vizcaino to the Yankees in June for Jim "The King" Leyritz. Leyritz didn't have much left in the tank, and he had fewer PA with the Dodgers than Hee Seop Choi before his major league career came to a close. Vizcaino was, well, Vizcaino. The Dodgers probably came out ahead on this deal, though, as Vizcaino was making a ridiculous $3.5 million in 2000 while Leyritz was only earning $1 mil. Sadly, I've been unable to locate the amount of money the Dodgers sent to the Yankees in the deal. Sigh.

Tuesday, March 01, 2005

THT Article Open Thread

...if you've got anything you want to say about this.

Is This A Dead Horse?

Imperabo, one of the Fourth Outfielder's favorite Primates, wrote this the other day during this discussion:

This is the one thing I do have concerns about with regard to Depo. He clearly puts a lot of weight on UZR (or something similar), and I'm afraid of it blowing up on us. The groundball staff won't be nearly as effective without good infield defense.

This is something I've considered as well. I thought I should take a closer look.

Let's assume the players at issue are Jeff Kent and Jose Valentin, since I don't think the Drew signing has seen much critique from a defensive standpoint. Firstly, how good do Kent and Valentin have to be at defense to be worth their contracts? If Kent is a +15 batter and Valentin a -10 batter (a low estimate, in my opinion), then with neutral defense they're +22 and -7 players when accounting for their positions. Baserunning makes those more like +20 and -5. Kent's getting an $8.5m average annual contract value and Valentin's getting $3.5m. For Kent, figuring $4m for an average player and $2m per extra win, his baseline is $8m, meaning he has to contribute 2.5 runs above average on defense to be worth the full $8.5m. For Valentin, since he'll sit a quarter of the time, his baseline is ($4m)*3/4 + (-.5)*3/4*$2m, which equals $2.25m, meaning he'll need to contribute 6 defensive runs above average to be worth the full $3.5m. At 3/4 playing time, that comes out to a rate of +8.

We'll deal with Valentin first. Valentin would have to be a pretty bad shortstop for anyone to figure him to be below average at third. Almost every defensive statistic likes Jose Valentin - He's been above average in David Pinto's Probablistic Model of Range in both 2003 and 2004. He has been above average in Davenports every season since coming to the White Sox in 2000. He has been above average in UZR in every season since 2000 except his -2 in 2001; he's also been at the top of the charts in UZR each of the past two seasons. He's been one of the best according to Studes' win shares calculations each of the past two seasons. While each of these methods has flaws, the fact that four different methods - including two using zone data and two not using zone data - show Valentin as being a very good defensive shortstop seems like a pretty strong preponderance of data. It's not just UZR that likes him.

Kent has certainly not been a favorite of most observers. However, we're not just dealing with UZR in his defense. Michael Humphreys' side-by-side comparison of UZR, ZR, Davenports, and Defensive Regression Analysis for 2001-2003 shows Kent as being above average by all four metrics, ranging from +2 to +12. Win Shares and Pinto's PMR were both down on him in 2003, but both had him near the top of the charts in 2004. That they were down on him in 2003 isn't terribly disconcerting, as the Davenport system has that as his only below average recent season (at -1). His 2004 Davenport was +10, and his 2004 UZR was +20. I also have Kent's 2004 ZR sitting around, and it's above average. Again, the consensus among several different methods seems to be that Kent is an above average fielder.

The other issue here is clearly one of misplaced expectations - most of us have seen that both Kent and Valentin dominated the 2004 UZR rankings and got excited. One season's worth of UZR, however, is not a sufficient sample size. A +20 over the course of one season doesn't mean that the player's true talent level is +20; it's probably closer to +5. Kent is probably in the +5 range defensively and Valentin at third is probably in the +10 or +15 range, with the caveat that both have shown an upward trend recently.

In summation, both Kent and Valentin look to be good bets to match their dollar value unless the available defensive evaluation systems are all pretty much junk. That's pretty good, considering that the average free agent signing is for above the player's actual value.

There's also another meta issue here. Paul DePodesta does not use any of these defensive metrics, or at least if he does he's using them to verify whatever proprietary methods he has going. If these signings (along with the Odalis Perez and Derek Lowe signings) signify that DePodesta's relying heavily on his defensive evaluation system, then we should be pleased since we have much evidence to indicate that his system is getting it right.

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