Thursday, March 20, 2008

A Rays Rebuttal

Baseball Crank chimes in with a 71-win projection for the 2008 Tampa Bay Rays, far off from PECOTA and myself.

BC makes some fair points regarding the Rays: Last year they set a league record for offensive strikeouts while leading the league in runs allowed by a wide margin. They can't really expect further improvement from Carlos Pena or B.J. Upton, and can't count on anything from Rocco Baldelli this year. How can a team like that improve by 20 wins?

Let me explain how I view the 2007 Devil Rays, our most convenient baseline for this year's projection. In 2007, Tampa did all this:

- Offensively, they posted above-average figures in both OBP and SLG, good for a team OPS+ of 103.
- Their pitchers led the league in strikeouts, with a better-than-average K/BB ratio and a groundball percentage of 43% versus a league average of 44%.
- Their defense sucked like Paris Hilton on spring break.
- The team scored 79 fewer runs and allowed 16 more runs than would be expected from their aggregate hitting and pitching stats.

Recapping the 2007 team:

- Offense: Above average.
- Pitching: Slightly above average.
- Defense: Maginot Line-esque.
- Luck: Witch doctor's curse.

Just how bad was the Rays' fielding? The Hardball Times says they gave up 142 more runs plays than an average defense would have. Baseball Prospectus pegs that figure at 71. Just by playing league-average defense this year, Tampa can cut their runs allowed by 70 or more.

Any objective look at the Rays defense this year gives a favorable prognosis. Carl Crawford and Jason Bartlett are elite defenders at their positions. Pena and Evan Longoria are plus gloves. The Rays will likely take hits on defense from second base, center field, and third base, but these shouldn't be enough to cancel out the surpluses. With Upton, Josh Wilson, and Brendan Harris permanently banished from the infield, the Rays should field an average team defensively this year.

To win 20 more games than in 2007, Tampa needs to improve their run differential by about 200 runs. Normally, this is a very tough task for a team that returns most of last year's regulars. However, Tampa has a huge head start: They expect to be 95 runs better in the luck department, and save 70 runs on defense. Now add in Evan Longoria, Matt Garza, and Troy Percival; subtract Jae Seo, Casey Fossum, and a parade of terrible relievers. The result is enough to balance out some regression from Pena and Upton, and then some.

Why am I so far off from Baseball Crank's conclusion? I don't believe the EWSL method is taking the proper steps to separate luck from skill in last year's results. If you think Andy Sonnanstine is actually a bad pitcher, rather than one who was simply victimized by his defense, you don't see a lot of hope for him. If you look at just runs scored and allowed, last year's Tampa squad looks hopeless. The component stats tell us otherwise, and that's why we go to the trouble of looking at them when formulating projections.


Dave said...

Hey Jon,

Those team fielding figures you quoted from THT are plays below average are plays below average, not runs.

j holz said...

fair. a play is worth almost a run anyway, and i used a conservative estimate for them, so this doesn't change my argument at all, but it's worth pointing out.

Donald said...

Jon, don't forget to mention the loss of Delmon Young and Rocco Baldelli in your analysis. That plus several young guys in the minors until June = sub-80 win season. Maybe next year?

Red Sox Talk