Depose Castro!

I will assume that most people reading this column are Twins fans. By extension, I will then assume that most people reading this column are upset that the Twins have chosen Juan Castro to be their opening day shortstop rather than Jason Bartlett, repeating one of the mistakes that cost them a semblance of an offense last year. Bartlett supposedly had a poor spring training, but based on the limited available spring training statistics, Bartlett out hit Castro and several other players who broke camp with the big team. Maybe it will only take a few weeks before Castro’s putrid bat will once again become untenable even at a position that does not require great offensive contributions. But why has it come to that? The front office’s continued inability to properly value their infielders’ offensive ability (or refusal acknowledge offense’s relative importance vis-à-vis defense) manifests itself in symptoms like Tony Batista and Doug Mientkiewicz. Castro just goes to show that they have not learned their lesson, and I am afraid that a lineup with this many sinkholes has no shot at the playoffs. Between choosing Baker over Liriano, Batista over Cuddyer (or anyone else who is upright and breathing) and Castro over Bartlett, the Twins have handicapped their meager probability of competing in an increasingly competitive division.

We do not know precisely the effect that playing Castro will have, but based on comprehensive statistical models, we can have a pretty good idea of the greatest probabilities. Defensively, both Bartlett and Castro are above average. Each player put up insanely high fielding Rate stats last year, Castro at 123 (Rate2) and Bartlett at 118, meaning they averaged 23 and 18 runs above average at the position every 100 games. We should temper our enthusiasm, though, as those numbers are so inconsistent with the rest of baseball that something has to be making noise in the machine. I suspect that Mike Cuddyer is to blame, as poor positioning or a general lack of skill produced putrid fielding rate stats for him at third the last two years (83, 89) commensurate with making good shortstops look like great ones by leaving more balls for them to play. Both players have career stats that make them look like good-very good defensive shortstops- Bartlett holding a slight advantage- but nothing like what they were last year. Combining their close results last year with Bartlett’s historical advantage, I believe neither can be considered a much greater asset at short, so offense should help make the decision.

And offense is a different story altogether. Based on PECOTA projections, Bartlett actually looks like a pretty good first option at short for most teams. With a .271/.339/.393 line, he adds on-base skills just above league average to his solid glove at the premium defensive position. To give some context, his .268 EQA is the same as Rafael Furcal, widely regarded as one of the top shortstops in the game and paid a salary worthy of his reputation. Castro, on the other hand, is gearing up for what could be a .254/.287/.378 year. Maybe that doesn’t look so much worse to you since the BA and SLG really are not that far off, but consider that a .050 point OBP deficit is like the difference between Cincy teammates Adam Dunn and Ryan Freel. Moreover, Castro shares his .237 EQA with Philadelphia futility infielder Ramon Martinez. “But Martinez would never play a full season,” you say. Exactly.

Running all sorts of numbers through Marginal Lineup Value formulas gives an idea of what Bartlett or Castro do to a lineup when you add them to it. Using the 2005 AL league averages of .268/.328/.424, a league average offense would score 785 runs. Plugging in the rate stats of Castro and Bartlett as everyday players, both offenses lose some runs, predictable with plus defenders at the left end of the defensive spectrum. Bartlett, though, with his decent ability to get on base, only costs the team three runs, factoring out to 782 runs for a complete season, which is quite good for someone with his glove. Castro’s contribution is not so optimistic, as the league average + Castro iteration would only score 762 runs. If defense is a push, which I maintain it would be, playing Castro over Batlett would cost the team at least two full wins, non insignificant in a tight division. In fact, the PECOTA projection for the AL Central have the Twins finishing four games behind the Indians for the division title, so two wins here and there would make all the difference. Looking at those same projections, we can see that the Twins project to 757 runs offensively, an unsurprisingly below-average figure, meaning that Bartlett would actually add offensive value whereas Castro would remain substantially below average.

I know this is not the note on which anyone wants to begin the season, but unfortunately, Batista and Castro do not strike such a harmonious chord.