Saturday, November 20, 2004

A Punless Hee Seop Choi Post

Now that I’ve gone back to rambling on about Jason Kendall, perhaps it’s also time I pontificate about Hee Seop Choi. The decision by the Dodgers to hold Choi out of the lineup in September and October was extremely disappointing, especially since I was already irritated by Jim Tracy’s use of him that resulted in his being removed from a lot of games in the sixth or seventh inning. I felt very strongly at the time that Choi should have been starting whenever there was a northpaw on the mound and that the Dodgers’ lefty heavy lineup should be distributed so that there was not a solid wall of left-handed batters for opposing LOOGYs to feast on.

Choi is turning 26 in March, and he’s already been marginalized by three organizations. Perhaps fortunately, there’s a good chance that the third one, the Dodgers, will prominently feature Choi next season.

Choi’s performance in his time with the Dodgers was, quite simply, terrible: .161/.289/.242. However, the reasons for that are quite identifiable. First, Choi had pretty bad luck on balls hit in play. Although his line drive percentage was down in his time with the Dodgers, this isn’t enough to account for his .227 BABIP. Using his hit ball data and JC Bradbury’s BABIP formula, I found that a more reasonable batting average on balls in play would be .284, which is still below average but much more reasonable. Second, Choi’s quality of opposition was pretty high. According to Baseball Prospectus’ data, the average batter would have hit .245/.327/.394 against the pitchers Choi faced while with LA, a revelation which shouldn’t surprise people who read this way back when. If you adjust his raw data for quality of opponent, his line becomes .173/.291/.260, but if this adjustment is made in concert with a BABIP adjustment his line bumps up to .216/.324/.325, which isn’t something that you should let play first base but which isn’t nearly as bad as his raw data would indicate. The third reason is that Choi didn’t hit any home runs with the Dodgers. Combining the BABIP data with an adjustment to convert 3 of his doubles into home runs (if he’d hit them at the same rate he did with Florida, he’d have ~3.3), he hits .242/.355/.435 with the Dodgers, and if we tweak that slightly again for quality of opposition (but this time use his Florida quality of opposition for SLG, since that’s where hit the home runs), we’re looking at .260/.358/.451.

Some of you are no doubt shaking your heads, thinking that what I’ve done in the previous paragraph is useless rationalization in an attempt to make a player I like appear better than he is. That is, to some extent, a valid reading, as it probably does characterize a chunk of my motivation. But at the same time, the death of baseball analysis is the refusal to consider alternate functions of data. Choi didn’t hit any home runs with the Dodgers, and there could be a lot of reasons for that—maybe Choi isn’t/wasn’t comfortable with the hitter’s backdrop there, maybe suddenly pitchers had learned to exploit the pundits’ perceived whole in Choi’s swing, maybe Choi and Tim Wallach aren’t a good fit, maybe the pressure from his being a PNG to Lo Duca-crazed Dodgers fans and his usage patterns changed his approach. However, it’s pointless to assume that Choi has lost his ability to hit balls out of the park forever simply because he had a bad month, and it’s pointless to evaluate his performance outside of the context provided by an analysis of his luck on balls in play and the quality of the pitchers he faced. I am not arguing that Choi did hit .260/.358/.451 but simply had bad luck; I am arguing that Choi continued to display nearly all the same skills that he’d displayed in his career through July 30, 2004, and that the one skill he ceased to display (home runs) was probably a sufferer from small sample size rather than a sudden decrease in ability.

So what is Choi’s outlook for the future? To try to answer this question, I took a look at a comparison suggested by Larry Mahnken: Carlos Delgado. Both Delgado and Choi are tall, fairly slow left-handed batters who specialize in power and walks. Delgado’s numbers in his age 24 and 25 seasons were, in many ways similar to Choi’s, with the only substantial differences being that Delgado hit home runs about 1.4 times as often as Choi and drew BB/HBP about .75 times as often as Choi. Those figures are based on major league performance, and I compared them to minor league performance. In the minors, Choi and Delgado hit home runs and drew walks with almost identical frequency. I also have made a subjective evaluation that Choi cannot be expected to hit as many singles as Delgado, who at least at this point in his career is a better line drive hitter than Choi and whose high BABIP in 1996 was somewhat more indicative of his talent than his average BABIP in 1997. There might be better comparisons for Choi, although my own examination of his list of PECOTA comparables from last season (Delgado was 8th most comparable at that point) didn’t reveal anyone whose skill set and performance more closely resembled those of Choi, but I think this one is pretty valid. So what would Choi look like if he followed Delgado’s development curve from age 26 onward? I sketched a composite of this “Chelgado” by taking Delgado’s numbers for each season and changing the following variables: his walks/HBP per plate appearance is increased by a factor of 1.33, his plate appearance per home run rate is increased by a factor of 1.5, and his batting average on balls in play is decreased by a factor of 0.89 (if you’re super curious about how I decided upon those figures, email me). Chelgado ends up with a median gross production average of .293 (age 29), with a high of .347 at age 28 and a low of .273 at age 32. His mean GPA for this seven-year period is .300, and even if we throw out Delgado’s ridiculously good age 28 and age 31 seasons the mean is a healthy .287. If you prefer runs created per 27 outs, Chelgado averages 7.02 or 6.366 dropping just the two highest seasons.

What’s more is that Delgado has been an everyday player his entire career despite having substantial platoon splits. If Choi were to follow Delgado’s curve while not starting against southpaws, it’s fair to raise those raw rate stats quite a bit. I don’t have Delgado’s career splits, but I ran a Chelgado simulation with his lefty/righty data from 2002-2004 and the GPA split is .254 to .315. I’m not certain how good of an idea it is for Choi to be permanently platooned, but I don’t think I’m opposed to it. I think there are enough right-handed hitters who can mash lefties and play at Choi’s end of the defensive spectrum for that to be a worthwhile strategy. For the Dodgers, this should be a relatively easy task, as Olmedo Saenz did a solid job as Jason Giambi’s platoon-mate in Oakland and his 2004 performance certainly indicates that that can continue. Heck, I wouldn’t rule out the possibility that Chin-Feng Chen could deliver performance above league average if installed as a southpaw specialist, although I admit to having no knowledge about his minor league splits. Given that a decent right-handed bat that can play first base is pretty easy to come by if you’re looking hard enough, I don’t think Choi’s terrible performance against lefties in his young major league career presents a significant problem.

While it’s conceivable that Choi will develop into a bona fide superstar, I think expectations should be more tempered. He probably won’t ever hit home runs or even singles as often as Carlos Delgado has, but it’s unlikely that all the power he demonstrated in the minor leagues has simply evaporated and won’t return. Furthermore, his outstanding and consistent ability to draw walks is almost guaranteed to stick around. I would argue that it’s more likely Choi will develop into a superstar than that he will be a below average first baseman. All told, Choi is a solid choice to be one of the ten best first basemen in baseball in each of the next six seasons, and I think that’s exactly what he’ll be. Given how little money he'll be making over the next four seasons, that looks like a pretty valuable commodity.

Wednesday, November 17, 2004

Putting the "Grrrrrrrrrrrrrrrrrrrr" in "Detroit Tigers" And the "???????" in "Washington ???????s"

This week, I've written entries defending decisions to acquire three high-priced players over thirty that might otherwise be looked upon with scorn. Three recent acquisitions have no such redeeming value, and I can't even conceive of how to defend them.

Jim Bowden's moves defy explanation. Well, okay, they don't, but they defy any explanation that would credit Bowden with a general sense of rationality, a perfunctory understanding of baseball, and a desire to succeed in his job. When Bowden was hired I considered posting something about what a stupid investment that would be for the 29 MLB teams that own the Washington Inpos, as no intelligent person would want to pay much for a franchise run by Bowden. I then realized, of course, that the owners would recoup that money because their teams would beat up on the fighting Bowdens. Let's just put it this way: if there's any logic to the Castilla and Guzman signings, Bowden is probably working the phones with Larry Beinfest right now because he wants a "proven run producer" in the outfield and the Marlins want to unload Juan Encarnacion. And yes, for those keeping score, Bowden actually did refer to Encarnacion as a "proven run producer" on ESPNews on July 30, 2004.

In fact, I've come up with the Inpo's new marketing campaign:


*caution: may not contain more than 30% baseball substance.

As Tom Waits said, "The large print giveth and the small print taketh away."

Meanwhile, Dave Dombrowski's Detroit decisions are deliciously D-minus-ly. Huzzah! Percival's 2.90 ERA in 2004 is surely a portent of things to come, yes? Well, no, as Percival's strikeout rates have been plummeting year to year and his home run rate isn't in great shape either. Toss in that he hasn't thrown 58 or more innings in any season since 1998, and six million dollars per season is getting pretty ridiculous.

Let me try to muster the strength to play devil's advocate on this one. Percival might not have the ability to overpower hitters any more, but his .230 opponents average was for real because, based on his batted ball type data, he induces a ton of fly balls, especially infield flies. I used one of my favorite little tools to find the answer here. JC Bradbury of The Sabernomics Blog did a study on the correlation between batted ball types and batting average on balls in play. I made a spreadsheet that adjusts a pitcher's stats based on his batted ball types and an average defense using Bradbury's findings and then inputs this data into the runs created formula to find RC and RC/27 as well as expected ERA ((RC/27)*((league R/G)/(league ERA))). Percival comes out with a 4.73 ERA, or 4.64 if adjusted for playing in Comerica. In other words, just over league average. So the Tigers are in line for 100 league average innings for $12 million, or $120,000 per inning. Deeeelightful. Hey, but I'd buy season tickets just to see an aging closer-- wouldn't you?

P.S. I don't know how in the world this is possible, but if you do a google search for "outfielder", my blog is #1 on the list, which also means if you forget my URL you can go to google, type "outfielder," and hit "I'm Feeling Lucky." I guess that's not bad for a blog that's on the young side of three months.

Sabean Strategy

Jon of Dodger Thoughts links to Bob Nightengale's latest piece. The sidebar for the column mentioned a rumored trade that would send Jason Kendall to the A's and then to the Dodgers, which I suppose was nice to see. But Nightengale's main claims about the Giants, particularly their interest in Steve Finley, are more intriguing.

When the Dodgers traded for Finley, I thought the move would be a strong one if Finley was platooned, with Jayson Werth playing in his stead against southpaws. As a lefty, he has substantial platoon splits but over the course of his career has always been an everyday player. Often, that's been pretty reasonable, because he's remained a decent hitter against LHP, especially in slugging, but he also contributes a lot of outs against southpaws. As it turns out, the Dodgers abundance in left-handed hitters with sizable platoon splits after the trading deadline made platooning Finley a non-option, and I suspect Finley's pride may have had something to do with it as well.

But the Giants are a different story. In my semi-defense of the Vizquel signing, I argued that the Giants were getting a commodity that would likely be undervalued because of platoon splits. What I wondered about but, for whatever reason, neglected to mention was that the signing would be a terrible one if the Giants didn't consult with Vizquel about his usage patterns before inking the contract-- did they ask him if he'd be willing to sit or move to 8th in the lineup on days when southpaws were on the mound? If Sabean didn't do that, then I agree with Mike Carminati that Sabean is "nuts."

The same should be done with Steve Finley. Finley's old and, despite the ridiculous Gold Glove voting, no longer even average defensively. But against RHP, Finley hit .282/.342/.507 last year and .290/.364/.502 over the last three years. Put that together with Marquis Grissom, whose RHP line was .266/.311/.404 in 2004 and .272/.303/.414 from 2002-2004 but whose LHP line was .315/.356/.577 in 2004 and .325/.369/.616 from 2002-2004, and you've got tremendous offensive production from centerfield, although no one's going to be too happy about the defense.

I can't say I agree with Sabean's overall strategy of consigning the future by signing an endless string of aged/aging free agents (especially his doing so when it costs the team a first-round draft pick), but if his signing of Vizquel and rumored pursuit of Finley are for the right reasons at least he's using reasonable tactics within that strategy. The 2007 Giants will almost certainly be terrible, but given the likelihood Bonds won't be around by then didn't we already know that?

Monday, November 15, 2004

Jason Kendall Redux

The NL MVP voting was released today, and the outcome at the top should surprise no one. A few things that I noticed, however, were further down the ballot.

For one, pitchers received very little support. This is perhaps rightfully so, but Randy Johnson and Ben Sheets would have been in my top ten. Why Roger Clemens and Eric Gagne received more votes than either Johnson or Sheets is no mystery, but in my book it certainly is faulty logic.

Furthermore, the two best players in Pennsylvania this season both received only marginal consideration, if any. Bobby Abreu, who, had he played in the AL this season, probably would have been second on my hypothetical MVP ballot behind Johan Santana, received only two votes, one for ninth place and one for tenth place. There were a lot of deserving candidates this year, and Abreu probably would have only been sixth or seventh on my ballot, but his lack of consideration is disappointing.

The other Pennsylvanian? Jason Kendall, who I named the best catcher in the NL this season. I would argue that in most seasons the top catcher in a league is probably among the ten most valuable players in the league. In 2004 in the NL, I don’t think that was the case, but you’d think at least the home-town papers would give him token ninth place votes. Alas, that was not the case as Kendall, like his fellow undervalued Pirates Oliver Perez and Jack Wilson, received no votes. Actually, only Johnny Estrada was the only NL catcher receiving any votes, a seventh place vote and two ninth place votes. Estrada had a much nicer story line—he was another epitome of Bobby Cox/John Schuerholz’ genius, a virtual unknown traded for a star who outperformed the star by leaps and bounds who started out the season strongly, while Kendall was represented in the media as an overpriced aging star unwanted by his penny-pinching non-contender team. However, Kendall’s season was, in my view, the superior one. Their offensive production was roughly equal, as Kendall’s .319/.399/.390 was good for a .280 EqA and Estrada’s .314/.378/.450 was good for a .286 EqA. However, Kendall played significantly more, logging 141.8 adjusted games at catcher versus 117.1 for Estrada. Furthermore, Kendall had an excellent season defensively while Estrada had a defensive season that equaled Kendall’s worst season as a major league catcher.

And yet Kendall is lumped into the group of “bad contract” players (Peter Gammons’ recent column is a good example). How bad is his contract? I’ll return to net win shares value here. According to the Hardball Times 2004 Baseball Annual, Kendall’s $8.57m contract was worth $4.8m, 87th best in major league baseball out of 763 contracts, putting him in the 88th percentile of value. Not too bad, huh? Kendall’s deal is worth 10, 11, and 13 million dollars in 2005, 2006, and 2007, respectively, which is indeed an awful lot of money to invest in a catcher on the wise side of thirty. But let’s contextualize that. For the deal to “break even” in net win shares value, Kendall would need to contribute about 25 win shares above replacement over the next three years.

I don’t have access to full win share data over his career, and the data I do have access to was calculated using slightly different formulas, so let’s just look at Kendall’s 2004. In 2004, Kendall produced 24.8 Win Shares, compared to his average baseline of 18 and his replacement baseline of about 13.5. His win shares were distributed as 18 batting and 6.8 fielding. Part of that came from his versatility; Kendall started 145 games and finished almost all of them, logging 1259 innings, while Jorge Posada was second in baseball in games started at catcher with 126 and Brian Schneider was second in innings with 1114. Over the course of Kendall’s career, that versatility has been a trend, as his AdjG has totaled 135 or more in all but two seasons since his rookie year. By comparison, Ivan Rodriguez’ career high in AdjG is 137.6, and he only was in the 130’s for a four-year stretch. Mike Piazza only broke the 130’s four times, as well. Kendall’s versatility is thus rare. It could be argued that this makes Kendall a high risk for fatigue, injury, and decline, but I’m not convinced of that. For one, Kendall’s splits both in 2004 and from 2002-2004 show a remarkable month to month consistency, and he’s actually peaked in September. For another thing, Kendall has no significant injury history aside from his infamous 2-year thumb injury; a Lexis-Nexis search didn’t reveal the cause of his thumb injury, but I’m willing to speculate that, it being a thumb injury, it wasn’t the result of stress put on his body by catching. So overall I’d say that there’s a very strong chance that Kendall can be counted on to catch 130 games over each of the next three seasons. Furthermore, the dearth of catching talent under the age of 30 available in baseball at this time leads me to believe that Kendall, who only turned 30 last June, is as good a bet as anyone to remain healthy and durable at the catcher position.

Defensively, Kendall has been all over the map in his career. Based on the rate stats at Baseball Prospectus, Kendall was a poor fielder as a rookie in 1996, a very good fielder in 1997 and 2004, an average fielder in 1998 and 2000, a subpar fielder from 2001 to 2003, and a brilliant fielder in limited playing time in 1999. His career mean Rate2 is 100, exactly average. Some would argue that his true defensive performance level going forward is probably along the lines of his 2001-2003 performance, and I think that might be fair. If he returns to that level, we could expect a performance roughly equal to Johnny Estrada’s this season. According to the eminently useful THT Annual, Estrada was good for 2.94 win shares per 1000 innings this season. If we prorate that over a somewhat conservative estimate of 1100 innings per season, Kendall projects at 3.2 win shares per season. However, an argument could certainly be made that Kendall’s 2001-2003 was spent dealing with and recovering from his thumb injury, and that either his career rate or his 2004 performance might be a better indicator of his performance moving forward. If we use his completely average career totals, that looks like Michael Barrett territory, 4.35 WS/1000 innings, which we’ll factor by 1120 innings per season for 4.9 win shares per season. And if we think Kendall’s defensive turnaround was permanent, we can maybe pencil him in for 6 defensive win shares per season, rounded down (way down) from his actual 6.8. Completely subjectively, I’m assigning a 40% probability to the conservative estimate, 45% to the moderate estimate, and 15% to the liberal estimate. That gives a weighted mean of 4.4 fielding win shares per season.
Offensively, Kendall has been fairly consistent. His EqA’s from ages 23 to 26 were in the .289 to .319 range, with a good chunk of that variability being from his home run rate fluctuation. At ages 27 and 28, his thumb injury hindered his effectiveness, and his EqA’s were .241 and .253, right around average for a catcher but not for one with such a hefty contract. In 2003 and 2004, his EqA rebounded to .286 and .280. How has his offensive performance broken down by segment? Here’s a chart:

1996 414 124 3 4 35 15 30 3 28 471 9.42 15.7 138 0.312 0.0909 14.786
1997 496 143 1 5 49 31 53 8 40 582 7.275 10.98 62 0.306 0.1311 12.4
1998 535 175 2 8 51 31 51 12 39 627 7.64634 12.29 44.583 0.338 0.1905 13.718
1999 280 93 0 4 38 12 32 8 23 334 6.68 10.44 35 0.348 0.2 12.174
2000 579 185 1 4 79 15 79 14 39 678 7.21277 8.582 41.357 0.348 0.1505 14.846
2001 606 161 0 2 44 20 48 10 24 672 10.5 14 60.6 0.275 0.1724 25.25
2002 545 154 0 2 49 9 29 3 28 605 10.431 20.86 181.67 0.293 0.0938 19.464
2003 587 191 1 3 49 25 40 6 32 665 8.98649 16.63 97.833 0.339 0.1304 18.344
2004 574 183 1 4 60 19 41 3 32 658 8.32911 16.05 191.33 0.336 0.0682 17.938

The data clearly illustrates three distinct periods in his career. The first was his prime, beginning with his second season, in which he had okay power, didn’t strike out much, hit a lot of line drives and ground balls and thus had a high batting average on balls in play, and drew a good amount of walks in addition to a ridiculous number of hit by pitches. The second period was his two-year thumb era, during which every aspect of his offensive play regressed except for his strikeout rate, which improved substantially. The third period is ages 29-30, in which his performance would be almost exactly identical if his home run totals were the same; during this period, he walks as much as he did during his peak and gets hit by pitches nearly as frequently, he still hits a ton of line drives and ground balls for singles, and his K rate is very low, but he has virtually no power. Furthermore, it should be noted that he fits the profile of a hitter who will hit well with runners in scoring position: he doesn’t strike out much and hits a lot of line drives and ground balls. If you asked me to predict his offensive production over the next three years, I’d probably reply that his production will look very similar to what it has been the last two years. I don’t think there’s any substantial evidence to indicate his ability to hit for a high batting average, avoid strikeouts, and draw walks and hit by pitches is set for a decline. I also don’t think there’s any substantial evidence to indicate his power will improve. As such, I’ve assigned conservative gut-level probabilities to his offensive win shares production so that there’s a 5% chance of collapse to 2001 levels (9 batting win shares per season), a 30% chance he’ll decline by 10% in each of the next three seasons (16, 15, 13 per season), a 60% chance he’ll do almost exactly the same as in 2004 (18 per season), and a 5% chance his power he’ll equal his 2004 performance level but with one amazing season thrown in over three years (two 18s and a 24). Add it all up and my arbitrary projection for Kendall’s total offensive win shares over the next season is 50.

Combining my projections for offensive and defensive win shares over the next three seasons, Kendall comes out at 63 win shares, or 21 per season. His replacement-level baseline in 2004 was roughly 13.5 win shares. Let’s lower that baseline to 12 since half of my projections included declines in production that would mean Kendall played less (all of the fielding win shares projections did this, and any substantial offensive decline would entail him receiving fewer plate appearances, especially since he hit leadoff all year for Pittsburgh in 2004). That means that these projections see Kendall contributing on average 9 win shares above replacement per season, and that’s a figure that is tempered by conservatism. Using the net win shares value calculator, that would make Kendall’s contract worth a net gain of $2.1m over the next three seasons, or $700K per season. I’m going to arbitrarily assume that market corrections will devalue that, however, and bring that figure down by $500K per season to a mere $200K per—still in the black. I think, therefore, that there’s a very strong likelihood that Jason Kendall’s contract will actually be a good one to have, straight up, over the next three seasons. Of course, Kendall could always lose a year or even his career to catastrophic injury, but he shares that risk with every other player in baseball. In fact, there’s also a substantial risk that I’ve been too conservative here; if we just increase his seasonal win shares above replacement by one (and I could give you dozens of reasons why we should), even with the half-mil market correction per season his contract is worth over a mil per year in net win shares value, which this year was about the 67th percentile of value.

The small handful of you who read this blog when I started may remember I posted the first part of a two-part Jason Kendall analysis when it seemed a remote possibility he could be traded to the catcher-needy Dodgers. Well, that particular rumor isn’t one I’ve seen burning in the hot stove recently, but it sure does make a lot of sense. The Dodgers’ minor league system will almost certainly not produce a decent, major-league ready catcher in the next three seasons. Meanwhile, the Pirates probably don’t have the resources to seriously compete anytime in the next two years. For each team, the marginal value produced by a trade would be somewhat maximized. The Dodgers are a team with a lot of organizational resources that can afford to “overspend” on a player of Jason Kendall’s caliber since, among playoff caliber-teams, value relative to average starts to become more important than value relative to replacement. The Pirates are a team that needs to turn a profit and sort through their farm system, and the uncertainty of when their young talent (especially their young pitching talent) will come together to be able to compete, especially with the organizational advantages owned by the Cubs, Cardinals, and Astros, makes payroll flexibility a premium. Furthermore, J.R. House probably won’t be a star, but he’s got a good shot at being a solid contributor in the major leagues (and the same could be said with a little less enthusiasm for Ronny Paulino). With the Dodgers having a quite deep crop of players in the low minors, I think Jason Kendall for, say, Jonathan Broxton and Willy Aybar would be a very good deal for both clubs. Whether the public relations part of such a deal would be manageable to the Pirates, I really can’t say, although the success of the Giles deal does, I imagine, give the Pirates some political capital—and maybe if the Pirates kicked a little dough LA’s way, LA could throw in Brian Myrow, who could look a lot like Jason Bay to Pirates fans (although I’m not 100% certain the Dodgers still have control over Myrow).

I know this is just annoying, restless hot stove analysis, but, well, as Bill Simmons would say, I feel very strongly about this.

Sunday, November 14, 2004


Was the Omar Vizquel signing as bad as it looked? Maybe not. First, let's look at Vizquel's 2004 season. Since this was a rebound year offensively for Vizquel, no doubt many predict a substantial decline next year. I'm not sure that I'd agree. Vizquel was not a recipient of extraordinary luck on balls in play, as his .317 batting average on balls in play is about right given his high 21.1 LD%. Furthermore, his 2004 season was, by segment, pretty much in line with the average of his last six seasons:

1999 664 10.0606 13.3 103 0.343 0.0909
2000 717 7.21739 9.96 88 0.31 0.0886
2001 693 11 9.63 79 0.277 0.027
2002 663 10.3594 10.4 143 0.296 0.1795
2003 285 9.82759 14.3 82 0.252 0.0909
2004 651 11.2241 10.5 97 0.302 0.1014

612 9.94818 11.3 98.67 0.296 0.0964

The segments there are Plate Appearances, Plate Appearances per Walks/Hit By Pitches, Plate Appearances per Strikeout, Isolated Power (BA - SLG), Batting Average on Balls in Play (for which I included sacrifices), and Batting Average on Balls Out of Play (HR/(HR + K)). His 2004 season fits pretty much right in line with the mean, and that mean would produce a .348 OBP and .365 SLG, good for a .248 Gross Production Average. By comparison, the average AL shortstop in 2004 had a .251 GPA and the average NL shortstop had a .240 GPA. And while it's not likely Vizquel will maintain the same level of performance he had over that period, chances are his decline will be gradual enough that his production will stay reasonably close.

Furthermore, many argue Vizquel's fielding skills appear to be diminished. I'm not sure that this is particularly accurate, either:

Rate2 ZR
1998 110 0.881
1999 91 0.862
2000 103 0.838
2001 109 0.84
2002 106 0.832
2003 113 0.892
2004 100 0.84

Vizquel's 2004 with the glove wasn't his best work, but given the year-to-year variance in Rate2 and Vizquel's steady zone rating, I think it's legitimate to argue that Vizquel will be several runs above average at shortstop over the next few years.

So Vizquel might be good, but is he worth all that cash? One way to evaluate that comes from the hard work of Studes over at the Hardball Times and his creation of Net Win Shares Value. Using Studes' calculator, I tried out a few models for what we can expect from Vizquel and his contract. Vizquel's 2004, in which he was paid, was worth about $503K according to Studes' calculator, as his 17 win shares versus his expected 16 win shares (i.e., what an average player would contribute with his playing time) were good for about 5 win shares above replacement level. Here's a few basic projections:

2005 19 16 4.25 3,439,457
16 15 4.25 1,541,076
2006 18 16 4 2,751,847
13 13 4 431,604
2007 16 15 4 1,697,191
10 11 4 -833,984

53 47 12.3 7,888,495
39 39 12.3 1,138,696

2005 14 15 4.25 -146,374
12 14 4.25 -1,201,030
2006 12 13 4 -201,190
10 12 4 -1,466,778
2007 9 11 4 -1,677,709
6 9 4 -2,943,296

35 39 12.3 -2,025,273
28 35 12.3 -5,611,104

So if Vizquel is significantly above average over the next few years, Sabean's made a very good deal. If he's pretty much average but plays with declining frequency, this deal is just about average, maybe even good. If he declines and is consistenly subpar, the deal is bad, but not stupefyingly so.

What all of the above analysis doesn't account for, however, is the particular situation Vizquel would be in in San Francisco. For one thing, the Giants' other shortstops, Deivi Cruz and Cody Ransom, are right-handed batters without significant historical platoon splits while Vizquel is a switch-hitter who's done much better against right-handed pitching (at least over the last several years), hitting .262/.311/.371 against southpaws over the last three years versus .282/.356/.400 against northpaws. Cruz' numbers against lefties over the last three years are about as valuable as Vizquel's, so if the Giants start Cruz against lefties they wouldn't derive a huge advantage but they would be able to keep Vizquel somewhat rested. So if Vizquel's plate appearances are distributed in the 4:1 RHP:LHP ratio of a lefty like Trot Nixon, his gross offensive production over the next three years could easily better what he did with Cleveland over the past several years. And if he were to be rested against LHP, then his defense could improve as well.

So let's try to take a stab at what Vizquel's production could look like versus his would-be replacements. Here's a chart I made based on somewhat arbitrary estimates of the productivity of Vizquel, Cruz, and Ransom, assuming that Vizquel sist against southpaws:


2005 Ransom 1 0.8 0
Ransom 6 5 421,863

Cruz 7 6 320,564
Cruz 16 14 1,375,220

Vizquel 14 16 2,173,870

22 19

22 24

2006 Ransom 8 6 0
Ransom 16 13 843,725

Vizquel 14 15 1,468,260
RLSS 6 4.5 0

22 21

22 17.5
2007 Ransom 8 6 -93,669
Ransom 16 13 625,164

Vizquel 14 13 -201,190
RLSS 6 4.5 0

22 19

22 17.5



I should note here that to make these calculations, I estimated Ransom's salary as $300,000 in 2005 and 2006 and then $650,000 in 2006 absent Vizquel's presence or $450,000 with Vizquel's presence since in the former case he would have earned more in arbitration. "RLSS" is, of course, a replacement-level shortstop earning the league minimum. The results slightly favor Brian Sabean's decision, but not by a very substantial margin. Keep in mind, of course, that there were a ton of subjective (though educated) guesses involved in these projections.

But to say that the signing is slightly better than just making do with Cruz, Ransom, and the mysterious replacement level shortstop is not necessarily to say that Sabean made the right decision. For one thing, all of the NWSV figures I used assumed 2004 contract values, and it's likely that the market correction will continue and that that would sap much of the marginal value from the Vizquel deal. Furthermore, Sabean made this signing in the context of a deep free agent shortstop pool, so arguably finding a better deal should have been easy, although I'm not sure of that argument since my cursory thinking on the subject indicates that the demand mostly matches the supply for free agent shortstops. Most importantly, perhaps, is that Net Win Shares Value is measured against average; I'm guessing that playoff teams will, outside of Club Steinbrenner, tend to average in the hundreds of thousands of dollars or even millions per position per season in NWSV. Meaning that, if my analysis above is a good rough approximation of the value to be earned by signing Vizquel, then the Giants have done something to make themselves competitive but have not made the kind of decision that makes the difference in whether a team makes the playoffs. And if the Giants end up with a bullpen next year that approximates what they had in 2004, the Vizquel signing will likely have been a mismanagement of resources that could have been spent finding superior relievers.

Of course, Sabean's backers will argue that he's got a great eye for veteran talent and that his more aged free agent signings have typically outperformed expectations. That might be true, but then again my analysis above already presupposes a bit of a rebound for Vizquel.

Overall, this signing is not, to me, an offensively bad one. But neither is this the kind of signing which indicates a team is playoff-bound. Paul DePodesta is Sabean's new competition, and DePo has a bigger budget and, at this point, a much better crop of talent in the minor leagues. The Barry Bonds advantage is huge, but it can only go so far. The time has come for Sabean to be creative, and I don't think uninspired Omar Vizquel signings are going to do it.

Saturday, November 13, 2004

Target Practice's list of the top 50 free agents leaves much to be desired, which I'm certain comes as no surprise. A few nuggets I picked out:
One more note: In my previous post, I discussed the Pirates agreeing to terms with Jose Mesa for a one-year, $2.5 million contract. What I didn't know at the time was that Salomon Torres was signed for $2.6 million... over two years. How on earth does 38-year-old Jose Mesa's 3.25 ERA/4.03 FIP in 69.1 IP top 32-year-old Torres' 2.64/3.48 in 92 IP such that he's worth twice as much? Well, there's this little stat called a "save"...

Thursday, November 11, 2004


Wednesday, November 10, 2004

Super World Cup

With the recent news that a baseball world cup is likely, I thought I'd take a look at how the teams might fare if composed of the best each country has to offer in major league baseball. My methodology was somewhat idiosyncratic, so I hope no one takes this as a major argument about the state of baseball or anything; it's just fun.

I only looked at the four countries that came close to fielding a full team of major leaguers, as Japan, Korea, Cuba, and Mexico don't have enough players in the major leagues for me to get reasonable results. That leaves the United States, the Dominican Republic, Venezuela, and Puerto Rico.

To look at the offenses, I used MLVr, which measures marginal run contribution over an average major league player. Each team's roster was constrained to 14 major league players and was constructed with defensive positions in mind, although I decided not to also study each team's defense. In coming up with the total runs above major league average, I used a distribution that treated each of the eight starters equally and credited the starters with 80% of the team's offensive output. The bench's production was weighted based on the better players receving the most playing time (if you're curious about the formula, email me). Here goes:
US Dominican Venezuela Puerto Rico
Johnny Estrada 0.171 Miguel Olivo 0.046 Victor Martinez 0.16 I-Rod 0.267
Helton 0.507 Albert Pujols 0.526 Miguel Cabrera 0.228 Carlos Delgado 0.179
Mark Loretta 0.289 Alfonso Soriano 0.029 Miguel Cairo 0.008 Jose Vidro 0.075
Alex Rodriguez 0.205 Tejada 0.24 Carlos Guillen 0.285 Dodgers 2B 0.082
Scott Rolen 0.419 Adrian Beltre 0.467 Melvin Mora 0.4 Mike Lowell 0.215
J.D. Drew 0.409 Vlad Guerrero 0.421 Bobby Abreu 0.365 Jose Cruz Jr. -0.031
Jim Edmonds 0.475 Jose Guillen 0.165 Endy Chavez -0.143 Carlos Beltran 0.243
Berkman 0.411 Manny Ramirez 0.363 Magglio Ordonez 0.126 Bernie Williams 0.05
Travis Hafner 0.398 David Ortiz 0.318 Ramon Hernandez 0.127 Javy 0.216
Thome 0.339 Rafael Furcal 0.031 Edgardo Alfonzo 0.033 Jorge Posada 0.191
Varitek 0.167 Carlos Pena 0.048 Omar Vizquel -0.011 Dodgers 2B 0.082
Eric Chavez 0.228 Aramis Ramirez 0.338 Omar Infante -0.008 Ricky Ledee -0.016
Aaron Rowand 0.23 Sosa 0.123 Cesar Izturis -0.032 Ruben Sierra -0.043
Michael Young 0.107 Alberto Castillo -0.069 Richard Hidalgo -0.04 Jose Valentin -0.083
2.63715 1.9975 1.17695 0.975

For Puerto Rico, I merged Jose Hernandez and Alex Cora into one player taking up two roster spots to make the accounting easier. As you can see, a few players are playing out of position. Oh, and in anticipation of the immortal one likely never participating and to give the other countries a fighting chance, I left out Mr. Bonds. Did anyone else realize that three catchers who've arguably been the best three of this decade so far are all Puerto Rican? Anyway, the US clearly has the best offense, although the Dominican's not looking so bad. What if it was just the US against the best of everyone else? I did that, too:

Johnny Estrada 0.171 I-Rod 0.267
Helton 0.507 Albert Pujols 0.526
Mark Loretta 0.289 Carlos Guillen 0.285
Alex Rodriguez 0.205 Tejada 0.24
Scott Rolen 0.419 Adrian Beltre 0.467
J.D. Drew 0.409 Vlad Guerrero 0.421
Jim Edmonds 0.475 Carlos Beltran 0.243
Berkman 0.411 Bobby Abreu 0.365
Travis Hafner 0.398 Melvin Mora 0.4
Thome 0.339 Manny Ramirez 0.363
Varitek 0.167 Aramis Ramirez 0.338
Eric Chavez 0.228 David Ortiz 0.318
Aaron Rowand 0.23 Javy Lopez 0.216
Michael Young 0.107 Jose Hernandez 0.263
2.63715 2.65185

Pretty even, huh? It'll have to come down to pitching. For this I used the Stuff metric available on Baseball Prospectus' DT cards, which might be my favorite quick evaluation metric for pitchers. I do somewhat regret using Stuff since a) it's not very well-known, b) it doesn't give much credit to pitchers who earn their high rates of outs on balls in play, and c) the distribution of correlation to runs allowed is a little funky. Here's what you have to know, from the BP glossary:
A rough indicator of the pitcher's overall dominance, based on normalized strikeout rates, walk rates, home run rates, runs allowed, and innings per game. "10" is league average, while "0" is roughly replacement level. The formula is as follows: Stuff = EqK9 * 6 - 1.333 * (EqERA + PERA) - 3 * EqBB9 - 5 * EqHR9 -3 * MAX{6-IP/G),0}

Once I got to working on Puerto Rico, I was suddenly unable to find major leaguers, so I gave them the benefit of the doubt and filled out their staff with replacement level pitchers. I also distributed the innings to come up with the overall staff Stuff average.

US Dominican Venezuela Puerto Rico
Johnson 45 Pedro 33 Johan Santana 44 Joel Pineiro 15
Sheets 40 Odalis Perez 11 Freddy Garcia 25 Javier Vazquez 13
Schmidt 36 Ramon Ortiz 2 Carlos Zambrano 24 Replacement 0
Peavy 32 Jose Lima -2 Kelvim Escobar 24 Replacement 0
Lidge 51 Francisco Cordero 25 K-Rod 45 J.C. Romero 9
B.J. Ryan 40 Armando Benitez 14 Juan Rincon 34 Kiko Calero 20
Joe Nathan 34 Juan Cruz 12 Gio Carrara 13 Replacement 0
Flash Gordon 27 Damaso Marte 8 Wilson Alvarez 12 Replacement 0
Foulke 21 Guillermo Mota 8 Uguie Urbina 8 Replacement 0
38.60377 13.15094 29.26415 7.641509

Johnson 45 Johan Santana 44
Sheets 40 Pedro 33
Schmidt 36 Freddy Garcia 25
Peavy 32 Kelvim Escobar 24
Lidge 51 K-Rod 45
B.J. Ryan 40 Juan Rincon 34
Joe Nathan 34 Francisco Cordero 25
Flash Gordon 27 Akinori Otsuka 23
Foulke 21 Mariano Rivera 15
38.60377 32.20755

It's not so close; the US really has a much better supply of pitchers. And after cursory research, I don't think there have been any years when the results would have been much different. This shouldn't be a surprising result since in the United States baseball isn't far from compulsory for athletic males in suburban settings, and thus US baseball has an enormous talent pool which has an extreme socioeconomic advantage over the rest of the competition and thus would be expected to have better coaching, nutrition, and access to training and learning resources. Despite this advantage, however, the US would be far from a shoo-in to win a Super World Cup due to (you know I can't go without saying this, right?) the extremely small sample size that would be involved. In any event, I'll probably be rooting for Team Greece to turn it around after their lackluster performance in Athens and win the cup.

p.s. I promise that sometime soon I'll work out the formatting in my tables so that they're not always this difficult to read.

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