« 2009 Hardball Times Season Preview | Main | Merry Christmas »

December 24, 2008

Who's Winning The Hamilton Volquez Trade

A few days before Christmas 2007, Texas traded Edinson Volquez, its most advanced pitching prospect, and reliever Danny Ray Herrera to Cincinnati for outfielder Josh Hamilton. My thoughts at the time:

I uneasily endorse this trade. That is to say, my anxiety about Volquez becoming a quality rotation figure is slightly higher than about Hamilton washing out. It’s a risky play for both teams. Texas just traded its third young and promising starter in three years, but Hamilton could become the best centerfielder in franchise history.

Hamilton and Volquez both had tremendous seasons. Hamilton led Texas in Win Shares and ranked 12th in all of baseball, and Volquez led the Cincinnati pitching corps. Even Herrera made his MLB debut, and though he didn’t earn any Win Shares, he did pitch much better than his unsightly 7.36 ERA.

As I’ve done previously, here’s an evaluation of the trade in terms of Win Shares and Wins Above Replacement:

Players
Win Shares
WARP
Hamilton
27
9.4
TEX Total
27
9.4
Volquez
17
7.3
Herrera
0
0.1
CIN Total
17
7.4
TEX Advantage
10
2.0

So, Texas enjoyed a net gain of about two to three wins (10 Win Shares = 3.3 wins). Victory, Texas!

Not so fast.

Because it’s winter and I have nothing better to do, I’ve explored how Texas would have performed had the trade not occurred. I did so by rationing Hamilton’s plate appearances to other Rangers, then stealing innings from other Ranger pitchers to account for the presence of Volquez. Then, I evaluated the gain or loss in value.

This is, of course, a fictional exercise. To keep the number of assumptions manageable, I’m only using players who actually took the field for Texas last season. Had the Hamilton-Volquez deal not consummated, Texas might have traded Volquez for a different outfielder, or for another pitcher, or kept him.

Hamilton Versus His Replacements

Hamilton entered 2008 with several questions, the least of which was the legitimacy of his’07 performance. More problematic were his health and former lifestyle. Hamilton completely quashed those doubts as a Ranger. Though he wore down a bit in the second half, he missed only six games all season and amassed 704 plate appearances.

I've assigned his 704 PAs to ten other players. Unless noted, all the performances in the extra PAs are simply extrapolated from real-life performances:

  • 200 to Nelson Cruz – Texas calls him up in July. I can’t fathom him equaling his actual 2008 line (.330/.421/.609) for another 200 appearances, so I averaged his 2008 performance with his career line (.251/.312/.431), giving him a still robust .291/.367/.520.

  • 114 to John Mayberry – Texas calls him up in August. I chopped 30 points off his AAA batting average, 40 from OBP, and 60 from slugging, giving him a line of .233/.276/.414. That seems reasonable. (Why 114 PAs? I needed someone to have a strange number so that all the players equaled Hamilton’s 704 appearances.)

  • 100 to Frank Catalanotto – Grudgingly, Texas hopes that Cat can revive his former .300-hitting semi-glory.

  • 100 to Brandon Boggs – Boggs receives a more generous opportunity.

  • 50 to German Duran – Duran played 30 innings of outfield this season. He gets another 100 or so minus Hamilton.

  • 50 to Jason Ellison – Ick. Someone has to spot Byrd in center after Murphy’s injury.

  • 10 to each of Marlon Byrd, Milton Bradley, Max Ramirez, and Jason Botts – Bradley plays injured slightly more often, Byrd gets less rest, and Botts and Ramirez enjoy a smidgeon of additional DH time.
In or Out? Player
PA
AVG
OBP
SLG
SB
CS
OUT Josh Hamilton 704 .304 .371 .530
9
1
IN Nelson Cruz 200 .291 .367 .520
5
2
IN John Mayberry 114 .233 .276 .414
2
1
IN Brandon Boggs 100 .226 .333 .399
1
1
IN Frank Catalanotto 100 .274 .342 .399
1
1
IN German Duran 50 .231 .275 .350
0
0
IN Jason Eliison 50 .231 .286 .231
0
0
IN Jason Botts 10 .158 .304 .395
0
0
IN Marlon Byrd 10 .298 .380 .462
0
0
IN Max Ramirez 10 .217 .345 .370
0
0
IN Milton Bradley 10 .321 .436 .563
0
0
IN TOTAL 704 .256 .329 .423
9
5

Hamilton batted .304/.371/.530 with nine stolen bases and one caught stealing. Driven mostly by Cruz, Hamilton’s ten replacements hit .256/.329/.423 with eight SB and four CS. That’s quite respectable, good for an OPS+ of 96, and above replacement level for an outfielder (even a corner). In terms of team-wide performance, losing Hamilton costs .006 in average, .005 in OBP, .012 in slugging, and one stolen base, and adds three caught stealing.

Translating that to a loss of runs can be calculated numerous ways, but I have my own formula based on a regression model that I’ve been using for fantasy baseball for several years:

Runs Scored Per Game = -4.87 + ( 25.58 * Average ) + ( 20.12 * OBP ) + ( 9.82 * Slugging ) + ( 0.26 * [ { SB – CS – CS } / Games ] )

(Note: The independent variables aren’t really independent, of course. Batting average is a significant subset of both OBP and slugging. Nevertheless, the model is quite robust in terms of several statistical tests. It’s impure, but it works.)

Rangers Offense AVG OBP SLG SB CS Runs / Game Total Runs
With Hamilton .283 .354 .462 81 25 5.61 909
Without Hamilton .277 .349 .450 81 29 5.41 876

My model predicts the Rangers would score 909 runs in 2008 (compared to the 901 they actually scored) with Hamilton. Without him, the prediction falls to 876, a loss of 33 runs, or about three wins.

I also attempted to estimate the effect of no Hamilton in the outfield. Advanced statistical systems rated him a poor centerfielder (where he spent most of the season) but an excellent right fielder. Replacing him with a combination of Byrd, Murphy and Boggs apparently benefits Texas by several runs, but they’re partially counteracted by having Catalanotto in left more often. Frankly, there’s too much noise in the ratings (Murphy is brilliant in right but terrible in left?) to reach a comfortable statistical conclusion about Hamilton’s absence. On the whole, I can’t imagine Texas faring better defensively without him. Call it a zero sum.

Volquez Versus His Replacements

How would Edinson Volquez have fared in Texas in 2008? First, not having Dusty Baker as manager reduce his 196 innings. That workload isn’t outrageous in and of itself; Volquez tossed 179 between Texas and the minors in 2007. On the other hand, he threw at least 110 pitches in 14 starts, one more than the entire Rangers rotation. Opinions vary on whether Baker warrants his reputation as an old-school arm slagger. Regardless, I believe Texas would have proceeded with more caution. Let’s say he throws 185 for Texas.

Second, Volquez would have allowed more runs. Texas plays in a tougher league and park, has an inferior defense, and, frankly, doesn’t have a strong history of developing starting pitching. So, to Volquez’s 82 runs allowed I multiplied the following factors:

1.03 for league
1.04 for park
1.02 for defense (Texas’s proportion of unearned runs versus Cincinnati’s. Texas and Cincinnati allowed similar hit rates on balls in play.)
1.10 for (lack of) development
0.94 for fewer innings pitched (185 versus 196)

Voila! Volquez allows 93 runs in 185 innings (4.52 Run Average) instead of 82 in 196 (3.76 RA). That seems reasonable.

I’ve also guesstimated the 185 innings that other pitchers wouldn’t have thrown. I suspect Texas still would have acquired Jason Jennings, providing an opening rotation of Kevin Millwood, Vicente Padilla, Volquez, Jennings, and Kason Gabbard, with Luis Mendoza biding his time in AAA. Here are the inning deductions:

  • 36 from Scott Feldman – Volquez’s presence delays Feldman’s debut and allows Texas to manage his workload better. Recall that Feldman started the season in AA to stretch out his arm in a benign environment but made only two starts before joining the Rangers.

  • 36 from Dustin Nippert (half his total) – Arguably, Texas doesn’t trade for Nippert with Volquez in tow, but I think not. If you can get a potentially salvageable big-league pitcher for a hard-throwing but erratic A-ball reliever (Jose Marte), you do it. That said, with Volquez around, Texas wearies of Nippert or hides him on the DL longer.

  • 32 from Luis Mendoza (half his total) – Mendoza doesn’t make the starting rotation or join Texas in April. He also gets less leeway (one would hope).

  • 28 from Sidney Ponson (half his total) –Notably, one-third of Ponson’s runs were unearned. He deserved worse than his 3.88 ERA..

  • 21 from Matt Harrison (one-quarter his total) – Harrison gets a little more AAA seasoning

  • 11 from Doug Mathis (half his total) – Mathis gets a briefer look.

  • 11 from Tommy Hunter (all) – Hunter doesn’t make the Majors in 2008.

  • 10 from Josh Rupe – Fewer fires to extinguish.

I’ve simply used each pitcher’s 2008 Run Average to estimate the number of runs allowed in those 185 innings versus Volquez:

In or Out? Player
IP
R
IN Edinson Volquez 185 93
OUT Scott Feldman 36 25
OUT Dustin Nippert 36 26
OUT Luis Mendoza 32 37
OUT Sidney Ponson 28 18
OUT Matt Harrison 21 14
OUT Doug Mathis 11 10
OUT Tommy Hunter 11 20
OUT Josh Rupe 10 6
OUT TOTAL 185 156
DIFFERENCE 63

The difference between Volquez and his replacements is an enormous 63 runs, about six wins.

Conclusions

Purely heads-up, Hamilton was worth two to three wins more than Volquez in 2008. In consideration of their alternatives, Hamilton was worth about three wins to Texas, but trading Volquez cost six wins. You might quibble with how I apportioned the replacements for Hamilton and Volquez, but the basic premise is clear. Hamilton displaced a group of hitters who performed at or above replacement level, while the Rangers who pitched Volquez’s allotted innings were well below replacement level. Though several pitchers on that list could improve substantially in 2009 and beyond, in 2008 they were a motley bunch. (It's also worth noting that some of Hamilton's heads-up advantage is due to Volquez's awful batting -- .098/.098/.098 with a 51% strikeout rate.)

The easy (and arguably legitimate) conclusion is that Texas foolishly traded pitching, a perpetual need, for hitting, which Texas has possessed in abundance. However, I don’t believe the Rangers expected the offense to be so good in 2008, nor the pitching so terrible. I certainly didn’t, though I did predict a 77-win season. They saw Hamilton as a huge upgrade, which of course he was, and more likely than Volquez to provide consistent value, which remains to be seen.

Interestingly, for a team that had a surfeit of quality outfielders for several years, the Reds have some gaping holes to fill. Gone are Hamilton, Adam Dunn, Ken Griffey (who batted a tepid .245/.355/.432), Ryan Freel, and even Corey Patterson, Cincinnati’s hilarious idea of a replacement for Hamilton. They do have Jay Bruce, but can Chris Dickerson continue to hit better in the Majors than in AAA? Who’s the 3rd outfielder? On September 21st in front of 22,000 mortified fans, Cincinnati’s starting outfield consisted of Jolbert Cabrera, Patterson, and Jerry Hairston Jr.

I think both teams would make this trade again. Cincinnati got an ace-worthy season from a player earning the league minimum. Texas benefited from its best season by an outfielder since Juan Gonzalez in 1999, a Home Run Derby performance that will be remembered for decades, and an amazing and genuine human-interest story.

Posted by Lucas at December 24, 2008 01:36 PM