Sean Rodriguez and the problem of poor fit

Several years ago, I read Charles Wheelan's excellent article on commencement speeches. As (at the time), a relatively recent college graduate, I'd become accustomed to accepting a lot of unsolicited life advice from well-intentioned Baby Boomers. With the exception of David Foster Wallace's beautiful 'This Is Water', it seemed to me that the commencement speech genre essentially constituted a string of trite exhortations toward realizing one's potential, ignoring, unsurprisingly enough, the socioeconomic quagmire in which my generation collectively floundered.

Wheelan's piece, in contrast, contained one line I'll never forget:

"You'll never read the following obituary: 'Bob Smith died yesterday at the age of 74. He finished life in 186th place.'"

The quote above doesn't, on its face, appear to have much to do with baseball. But Wheelan's words do remind us that our attempts to quantify the concept of 'value' as it relates to the endeavors of human beings are inherently an imperfect exercise. This reminder is especially prescient for those who, like the present author, spend an inordinate proportion of their lives trying to glean value from statistics.

It's become popular, with the increasing acceptance of data analysis in professional sports, to try to express players' respective values via a single, all-encompassing measure of worth. WAR, for baseball, RPM or (less usefully) PER for basketball, QBR for quarterbacks. These measures have varying degrees of quantitative validity, but the general underlying philosophy is the same--they are based on a belief that we can objectively measure the value a player provided to his team, and by doing so compare players to determine relative value. The proliferation of all-in-one stats is driven by analytical necessity (is player one better than player two? By how much?), but that doesn't mean it's an exact science.

For one, players don't exist in isolation. I imagine it's relatively uncontroversial to state that for basketball and football, players' value is conditional on the context in which they're placed. However advanced the ridge regressions employed by RPM might be, you can't really evaluate anyone on the Cavaliers in isolation from LeBron James. And nobody will ever know how Wes Welker might've performed without Tom Brady and Peyton Manning.

What's less obvious, however, and more important to the point of this article, is that the same phenomena are at work in baseball. If you employ Manny Ramirez (or, more recently, Hanley Ramirez), the value of a defense-first fourth outfielder goes up. If, on the other hand, your starting outfield includes Starling Marte, Andrew McCutchen, and Gregory Polanco, having a defensive specialist outfielder on the bench makes little sense. All of this is a long way of saying that we need to account for how players' unique skills fit together both when assembling a team and when retrospectively evaluating player performance.

Consider Sean Rodriguez. While the Serpico lookalike's athleticism, versatility, and ridiculous diving catches have endeared him to both fans and Clint Hurdle, advanced statistics are less kind, pegging him at -0.6 WAR thus far. Which is awful.

This is a case, however, when the all-in-one metric might not be perfect--WAR needs a bit of context. If we navigate to Rodriguez's FanGraphs page, we can see that a large portion of this negative value (5.5 runs below average) is driven by his .232/.252/.357 batting line. Understandable, given that his plate approach this year is basically "Screw it, I'm throwing some haymakers" (see below).

Prior to this year, Sean Rodriguez was a non-awful hitter. The projection systems ZiPS and Steamer both expect him to return to being a non-awful hitter, forecasting .238/.293/.385 and .237/.286/.382 batting lines, respectively. I'm a bit bearish on his bat, but I don't think it's unreasonable to believe that Rodriguez will be a better hitter moving forward than the guy we've seen thus far.

The more interesting part is that Rodriguez, who to the naked eye looks like a defensive savant, has apparently been worth 3.6 runs below average on defense. 1.7 of this is for fielding skill, and 1.9 of it is positional, which means a penalty he's incurred for playing positions further down the defensive spectrum (e.g., first base, right field).

Here are Rodriguez's career defensive metrics:

And here are his metrics for 2015 (note that the fan scouting report is not yet available):

Rodriguez's value primarily comes from two things: playing good defense and preventing playing time from being allocated to truly awful players. But it's tough to play defense well enough to accrue positive defensive value when playing first base and the corner outfield positions, and WAR does not assign any value to multi-positional versatility. Especially considering that while Rodriguez is a plus defensive infielder, it's not clear that he's quite as proficient in the outfield.

The situation isn't entirely the fault of the Pirates or Rodriguez--the Pirates didn't know, when they acquired Rodriguez, that they'd later have the opportunity to add Jung-Ho Kang to effectively fill Rodriguez's role. And it's not really Rodriguez's fault that the Pirates don't need him to play second base. But there's no denying that the resulting situation is suboptimal--Rodriguez is spending a bunch of time playing first base and the corner outfield positions, and he doesn't really have the bat to do that anymore. And he's spending relatively little time playing other infield positions, even though he at least superficially appears to have the glove to do so.

Rodriguez is a player whose skills, while useful, aren't perfectly aligned with the role he's ended up playing. And the thing is, he's probably the best fit of any of the guys on our bench.

Last year, we were all frustrated by the Pirates' apparent insistence on spending several roster spots on no-hit backup middle infielders (Nix, Morel, Martinez). It wasn't simply that each of these players were individually useless (although that was also the case)--it was that their skill sets were almost totally redundant, and consequently we ended up using some of the worst hitters in the major leagues in high-leverage pinch-hitting situations.

This year, the team seemed to be taking a different approach, judging by Andrew Lambo's inclusion on the Opening Day roster ahead of defensive specialist Pedro Florimon. But ineffectiveness and injuries to Lambo and Corey Hart, combined with the trade of Travis Snider, have left us with Steve Lombardozzi, Gorkys Hernandez, and Travis Ishikawa joining Sean Rodriguez and Chris Stewart on the bench. Each of these three guys, individually, is probably better than any of the Nix/Morel/Martinez trio. But that isn't really the point. The point is that their combined skill sets don't add anything to the other ten position players on the roster.

Post-Harrison injury, here are the different roles our bench needs to be able to fill, with the player most able to fill that role in parenthesis:

Backup catcher (Stewart)

Pinch-hitter (Rodriguez)

Utility infielder (Rodriguez)

Polanco platoon buddy (Rodriguez)

Alvarez platoon buddy (Rodriguez)

Alvarez defensive replacement (Rodriguez)

See the problem? I am fully prepared to believe that Gorkys Hernandez is an excellent defensive outfielder, but unless Starling Marte is going to miss significant time we have roughly zero use for a defensive sub in the outfield. We do have quite a bit of use for a right fielder who can hit lefties, but there's little reason to believe that Hernandez has anything approaching an MLB-caliber bat or should ever really be starting a major league game.

And while Steve Lombardozzi is capable of doing lots of different things, he doesn't project to do any of them well enough to be an asset--he's basically a worse version of Rodriguez. Lombardozzi has primarily played second base during his brief major league career, but he grades out as a below-average defender at the position, most recently accumulating -2.8 UZR in just 160 innings at second base with the Orioles last year. Nor do the Pirates appear interested in his defense or positional versatility, given that he has yet to see a single inning in the field. It's unclear what, exactly, the Pirates expect Lombardozzi to contribute--his current role means that he'll see virtually all of his playing time as a pinch-hitter, but he's a terrible hitter.

Ishikawa is a bit more intriguing--despite projection systems' pessimism regarding his bat (collectively, .231/.294/.355) he was an average major league hitter as recently as last year. He's also left-handed and a fantastic defensive first baseman (13.2 career UZR/150), which means we'll probably still be able to remove Pedro from games while Rodriguez is off doing the nineteen other things we need him to do.

But the issue stands that none of our bench guys can really hit all that much, and none of them is well-suited to platooning with Polanco. This sort of roster construction is how, with the go-ahead run on second base in the eighth inning of a tied game, you end up with Steve Lombardozzi, a career .264/.296/.338 hitter, at the plate.

It's what made Jose Tabata's DFAing puzzling to me. Tabata is an easy target--he's not a particularly great player, he's 5253479 years old, he's overpaid, and he has a terrible neck tattoo. But he's also pretty easily a superior offensive player to our current pinch-hitting options, and he's probably a better option to platoon with Polanco than Hernandez is. It's not necessarily that Tabata is a better player than Hernandez, it's that his skills fill a hole for us, while Hernandez's duplicate skills that other players already possess. And if they really wanted to demote Tabata, why not bring up Jaff Decker, who projects as a much better hitter than Hernandez and could also pinch run and play both outfield corners credibly?

This isn't a huge deal--if your biggest area of concern is the last couple guys on your bench, life is pretty good, and complaining seems a bit like being the sorority girl who melts down because the Starbucks barista misspelled 'Kaitlynn' on her cup. It does, however, speak to what seems like a potential blind spot of our otherwise-excellent front office--for the second year running, we've devoted several bench slots to guys who don't have any apparent role on the team.

Maybe they believe heavily in chemistry (I do), and think that Lombardozzi is a positive influence in the clubhouse. Maybe Tabata and Decker keyed Huntington's car. Who knows? But right now a very good front office is doing something that doesn't appear to make much sense.