Replacement Level Yankees Weblog

Disclaimer: If you think this is the official website of the New York Yankees, you're an idiot. Go away.

See The Changes
by Larry Mahnken

Since starting this blog 9½ months ago, I've suggested several times that the Yankees trade Alfonso Soriano. This isn't because he's a bad player; indeed, he's a very good player. In 2002 he was the 10th most valuable offensive player in MLB, and in 2003, he was 15th. But when you evaluate him solely as a hitter, you see that there's about 40-50 better hitters than him. That's still very good, but the perception created by Soriano's Home Run power and Batting Average is that he's one of the 15 or so best hitters in the game, though his dreadful postseason had to have damaged that perception somewhat. Still, that's a sizeable gap between perception and reality, and there is where the opportunity for a team to improve itself exists.

This past offseason, I advocated two rumored trades involving Soriano. I called for the Yankees to pursue a Soriano for A-Rod deal last autumn and throughout the Boston affair, feeling that if Texas was foolish enough to part with Rodriguez, that Soriano would probably be enough to land him. As it turns out, Soriano was just about enough to land A-Rod, though I had assumed the Yankees would have had to take all of Rodriguez's contract to make the deal work.

The other trade I advocated was a trade for Carlos Beltran, which was part of a bold plan I had conceived to refashion the Yankees as a better hitting team, a better pitching team, and a better fielding team. For this plan, I was ridiculed, and criticized by some for making proposals that involved the Yankees throwing their financial weight around and signing some of the best free agents. As it turns out, the Yankees did spend money liberally, they did act boldly, and indeed, they went further than I had advocated in assembling the 2004 squad. The 2004 Yankees will be a better hitting, better pitching, and better fielding team than the 2003 team, though the actual plan they followed was almost entirely diffent from mine.

Being sort of right on these two things isn't much of a reason to crow: I was wrong on quite a bit of it, and I'm way off base often enough to even out the times I'm very right. I didn't call for the A-Rod trade because I'm a genius, I did because I was able to put aside idealism and see the situation as it really was, more or less, and having done that, the trade was obvious. Similarly, I called for the trades of Soriano and Nick Johnson, and an agressive pursuit of free agents because I could see that the Yankees couldn't improve very much, and would still be left with glaring weaknesses if they weren't willing to spend big, and part with their young stars. As great as Nick Johnson is going to be, he's still a first baseman, and they have one of those already. In trading Johnson, the Yankees were able to (hopefully) move Bernie out of center, and accquired one of the best pitchers in baseball to fill a massive hole. Most of you guys would have come to the same conclusions as me when you put aside personal loyalties and saw the players in terms of their utility and value. They were just too obvious.

So, I won't give myself props for being sort of right, because I know I don't deserve it. But unfortunately for the rest of you, this does prove I'm not a total nitwit, so I'll keep throwing my opinions out there, and I won't care quite so much if anyone else agrees with me.

I've got to say, I'm ecstatic about this A-Rod trade. Using Baseball Prospectus' PECOTA projection system, I compared the Yankees' lineup possibilities for this season to what last season's lineup would be expected to do this season, using EqMLVR¹ to measure how many runs they'd score above an average lineup, though I used UZR to measure defense. (If anyone from BPro is reading this, and this violates the request to not distribute the PECOTA spreadsheet, email me and let me know what level of information you're okay with me sharing--I'll make a change).

           2003 Lineup in 2004²                            2004 Lineup³
Ps Name             EqMLVR Defense Total    Ps Name             EqMLVR Defense Total
2B Alfonso Soriano   .161   -.025   .136    CF Kenny Lofton     -.002    .074   .072
SS Derek Jeter       .092   -.191 -.100    SS Derek Jeter       .092   -.191 -.100
1B Jason Giambi      .327   -.025   .303    3B Alex Rodriguez    .332    .062   .394
CF Bernie Williams   .111   -.198 -.087    1B Jason Giambi      .327   -.025   .303
LF Hideki Matsui     .098   -.025   .074    RF Gary Sheffield    .245   -.136   .109
C Jorge Posada      .108   -.025   .083    DH Bernie Williams   .111   ---   .111
DH Nick Johnson      .206     ---   .206    C Jorge Posada      .108   -.025   .083
3B Aaron Boone       .007    .031   .038    LF Hideki Matsui     .098   -.025   .074
RF Juan Rivera       .007   -.019 -.011    2B Miguel Cairo     -.064   -.043  -.107
Total               1.117   -.475 .642    Total               1.247   -.309   .939
162 Game Total      181.0   -77.0 104.0    162 Game Total      202.1   -50.0 152.1

It goes without saying that the Yankees' bullpen is better than it was last year, and the rotation is arguably better, too. Even if that was a wash, the new Yankees' lineup rates as being better than the old lineup would have been by 5½ wins. But the Yankees suffered injuries last season, and several players had down years. Their actual MLVR was .914--148.1 runs above average over a full season. If you assume that the Yankees will meet these projections and stay healthy, (and that their defense was 77 runs below average last season) then their new lineup is almost *9 wins better than last season*. Factor in improvements to the rotation and bullpen, and the Yankees can reasonably win 100 games even with some injuries.

And then there's the question of who's on third. Using Tango's Positional adjustments (link below), I ran the same lineup with Jeter at third and A-Rod at short instead, and the difference was about eighteen runs. That's almost 2 wins. It can't be emphasized how foolish a move it is to put A-Rod at third, but every comment from A-Rod and Torre indicates that there is pretty much no chance of it happening any other way. But at least I have something to complain about all year...

Another thing that's being said is that the A-Rod trade will stop the Yankees from signing Travis Lee. Perhaps it will: Brian Cashman has said the Yankees aren't adding another penny to payroll this offseason, but perhaps he considers Lee an almost done-deal. The motivation for making that signing remains the same, A-Rod's accquistion doesn't change it at all. Jason Giambi's knees are still a concern, and Travis Lee is a much better option to start at first when Giambi DH's. I ran the PECOTA numbers for the Yankees with Lee in the mix, and found that with Lee in the lineup instead of Bernie and at first instead of Giambi, the Yankees are 4.9 runs better overall, and 17 runs better on defense. Put Bernie in left and bench Matsui, and the Yankees are 2.9 runs worse than that, and 5 runs worse on defense. In other words, This signing of Lee makes Bernie theoretically expendable, but the difference is statistically insignifcant. However, the upgrade in defense by having Lee in the lineup over either Bernie or Matsui means that one of the latter two should sit.

The last thing I looked at when evaluating lineup possibilities is second base, particularly defense. I ran the lineup with Jose Vidro in it to see how that would be (2½ wins better), but the main thing I wanted to see what how the Yankees could improve themselves defensively. To give themselves an average defense without Travis Lee, the Yankees would need to add a player who is 22 runs above average to replace Miguel Cairo (7 runs below) at second base. No such player is available, and all of the best defensive second basemen are unavailable to the Yankees, with the exception of perhaps Craig Counsell. With Travis Lee, the Yankees' defense gets above average with Counsell. Of course, having an average defense isn't anything special in and of itself, and sacrificing 2 runs of offense for 1 run of defense is a bad idea, so the Yankees should explore all options.

Now, almost everything I just said was based on projections of numbers based on numbers, and while those projections are pretty darn accurate as far as projections go, they're still just projections, and not things that have happened. I usually don't rely on numbers this much for my analysis (at least, I don't think I do), because they shouldn't really be used as the basis of a point, but rather be used to illuminate one. The point they do illuminate is this: the Yankees are better now than they were a year ago, they're better now than they were before the A-Rod trade, they're better with A-Rod and Cairo than they were with Soriano and Boone, and they're better with A-Rod at shortstop and D.J. at third. How much better they are is debatable, but the fact that they are is indisputable.

There are a lot of misconceptions out there about what this trade does. One is that the Yankees gave up too much for A-Rod, that the increase in payroll is not worth the increase in value. Alternately, some have said that the Yankees are only increasing their payroll $850K to bring in A-Rod, which is wrong to the other extreme.

Drew Henson was due $4 MM before going back to football, Aaron Boone was due $5.75 MM before injuring his knee, and Alfonso Soriano is getting paid $5.4 MM in 2004. With Texas picking up a large chunk of A-Rod's salary, A-Rod will be getting paid $16 MM by the Yankees this season, only $850,000 more than the $15.15 due to Henson, Boone and Soriano.

But the release of Henson and imminent release of Boone would have happened regardless of this trade, and Boone will have to be paid about $920K when released. I suppose that if one compares the Yankees now to from before the Boone injury, you could say that the Yankees have increased their payroll $1,767,553 and improved by 1.8 wins, but the Yankees are really paying $10.6 MM for it.

Which is a bad deal, but this trade doesn't exist for just this year. Soriano's salary is going to go up in arbitration in 2005, and up again--probably over $10 MM--in 2006. Then, in 2007, Soriano is a free agent and might make as much as A-Rod, while A-Rod is with the Yankees for another four seasons. The Yankees are certainly paying millions more for A-Rod than Soriano, but A-Rod will probably be worth about 20-25 wins more than Soriano over the next seven years. It's a good deal.

Another idea going around is that Soriano is almost as good as A-Rod, could become almost as good as A-Rod, or perhaps is even better than A-Rod. If you fiddle enough with the numbers, you can find some way to show that Soriano is better than Rodriguez. Soriano has more home runs on the road than A-Rod in the past two years, and more doubles. He has a higher Batting Average on the road than A-Rod, and a higher Slugging Percentage, too! It becomes obvious that Soriano is just as good as A-Rod, maybe even better, and that the difference between them is a product of their ballparks. Put A-Rod in Yankee Stadium and Soriano in Arlington, and you'll see who's better.

Well, those splits contain a whole lot of noise. Yes, Soriano has more doubles and homers on the road than A-Rod, but that's in 105 more ABs. Yeah, his batting average is higher, but his On-Base Percentage is 32 points lower. His slugging percentage is higher, but by only 13 points. Soriano's GPA is eleven points lower than A-Rod on the road, which is fairly close, but then, there's still more noise.

A-Rod's road games don't include The Ballpark in Arlington, Soriano's do. Soriano's road games don't include Yankee Stadium, A-Rod's do. Half of Soriano's road games were in AL East Stadiums; Hitters' parks Fenway and SkyDome, pitchers' park Camden Yards, and neutral park Tropicana Field. A-Rod, on the other hand, has played a third of his road games in AL West parks, all of which are pitchers' parks.

Road stats are not neutral stats. If you want to adjust for park, you don't take away the home games, you adjust the numbers for a park factor, which tells you what the player would have done if they played an equal number of games in every park. You compare OPS+, or EqA, or MLV. You don't take road stats. And the park-adjusted stats say that A-Rod is at least 2½ games better than Soriano, perhaps as much as 4 wins.

Finally, could Alfonso Soriano become as good as A-Rod? Well, I guess maybe he could, but that would involve an extreme improvement in his plate discipline. I think it's much more likely that Soriano will collapse than become as good as Rodriguez, but more likely than either that he's reached his level. Yeah, he's only 26, but he declined last year from his 2002 performance. There is little indication in his past performance or style of play that he'll take a big step forward in the next couple of seasons.

People are jumping through hoops to make this trade look bad for the Yankees because most baseball fans are really really sick of the Yankees winning all the time, and there's nothing people would like better than to see the Yankees lose. Fans are making these arguments because it makes them feel better about the trade, sportswriters are making the arguments because that's what their audience wants to hear. But it wasn't a bad trade for the Yankees, it was a good one. There's no honest way to make it look like it wasn't.

¹ EqMLVR is a context-adjusted measure of how many additional runs a lineup would score with the player added to it. I divided UZR runs, which are compared to the league average, by 162 games so they would be on an equal scale. .000 EqMLVR means a player is an average offensive player, and is not adjusted for position, so Derek Jeter's overall value of -.050 doesn't mean he's a below average shortstop, because the replacement level for shortstop is lower than at other positions. Further, since run scoring in baseball is not linear, EqMLVR is not an exact estimate of the runs a team would score, since a good hitter is worth more to a good hitting team. In other words, if all the players listed met their projections, the Yankees should outperform the overall projection.

² I put Juan Rivera in the 2003 version of the lineup because PECOTA projected him as a better player. Not having UZR for Rivera, I used BPro's defensive value projection. Don't worry about the lineup order, it didn't factor into these calculations.

³ Alex Rodriguez's defensive value at third base was determined using TangoTiger's True Talent Fielding Level Ratings.

This post was edited due to an error when I entered UZR onto my spreadsheet. The Yankees actually improved a game more than I had originally calculated.

--posted at 12:00 AM by Larry Mahnken / |

Larry Mahnken and SG's
	Replacement Level Yankees Weblog
"Hey, it's free!"