What kind of player, really, is Rick Ankiel the center-fielder? What value would it present to the team if Colby Rasmus turns into the Next Big Thing in that position, instead? What about Jay, or Pham, in the future? What makes a superlative center-fielder, or even an average one?
In search of something more accurate than this year's WAR charts, which are a counting measure and only take one year's defense into account, I followed a methodology similar to that in my recent posts on second base (details of methodology below).
First, let's define the mean, putting some names and faces to the numbers.
Considering the CFs who spent regular time in CF in 2008, the median UZR/150 was 1.0 run above average. League average defense was played by:
Anderson, Brian (2.1)
Converting wOBA into runs over a season (formula below), the median of the CFs was 2.2 runs above average (the arithmetic mean was pulled down by a few truly abysmal performances). The league average CFs, by wOBA, were:
Johnson, R (3.1)
Young, Chris (-0.6)
Overall (adding offense and defense together) the median was -0.2 runs below average. The overall league average CFs for 2008 were:
The cream does rise to the top, and that's what we're presumably hoping for with Rasmus and Ankiel.
The top defensive CFs were (with runs saved above average):
Jones, Andruw (13.6)
The top offensive CFs were (with runs produced above average):
Taking offense and defense together, I looked for a clear break in the data, some point of discontinuity between consecutive results that would show the difference between the really good and the average CFs. One emerged, at 9.8 runs above average, with a jump down to 5.6 runs above average for the next best player. Here are the CFs who really performed.
Some observations gleaned from this data:
1. Primo center field is a young man's game. Of the top twelve overall CFs, ten have not made it past arbitration. Perhaps this is a wave crest, resulting from the decline of players like Jones, Edmonds and Hunter. Don't blame my defensive weighting system, because it hurt more of the young guys than it helped. Of the eight contracts valued at over $5MM per year, as many were in the bottom twelve as in the top twelve.
2. Rick Ankiel gets into the top group by benefit of the doubt. The weighting system I used defensively pulled him up toward league average. If he really is a -8 defender (and there is no way to know right now, that I am aware of), he belongs in the league average group. For now. He's 30. See point #1.
3. Perhaps Colby Rasmus' young years, therefore, are the ones really to value. Maybe we are best getting him out there now, because his legs will be beat to death by age 28. His CHONE wOBA projection is league average for this year (0.330, which is a chunk lower than he has shown in any healthy season at any level). That means he will be worse offensively than Coco Crisp was last year, which I just don't believe.
4. According to Dave Cameron, the true value for freely available talent in CF is only -5 runs below league average. This means that 13 CFs who received considerable playing time in 2008 (and some BIG contracts) played at least 5 runs *below* the real replacement level. The high replacement level value for CFs was due to an abundance of good-glove, no-hit, minor-league free agents.
(Possibly relevant link: http://www.fangraphs.com/blogs/index.php/minimum-acceptable-offense/)
5. Unfortunately, Jay is a lefty, too, and older than Rasmus. However, "older" only means 24. Maybe raise the expectations for young CFs to make the team? But, what do you do with him, if Raz fulfills expectations?
6. Maybe it goes without saying, but be very wary of signing Ankiel to a long-term contract and expecting him to play CF (maybe b/c Colby was traded away?). Better to look to the young guns = cheap guns = better than most of the guys you can pay to sign.
The players surveyed are those who appeared in at least 80 games in CF last year, according to their Fangraphs defensive chart. All statistics are from Fangraphs. Defensive stats are only from the players' time in CF.
I regressed the players' 2008 wOBA toward their career average wOBA, using a 3 to 1 weighting (3*2008 + 1*Career Average)/4. Then I converted the wOBA to runs against average over a season, using the following formula: ((wOBA-0.330)/1.20)*600.
For defense, I looked for a three-year total for UZR/150. This was a pain, in that so many young players were involved. To produce a more reliable estimate, I regressed players with no data or insufficient data for a year toward league average. The weighting was 6, 4, 3, for years 2008, 2007, and 2006. Everyone had data for 2008, and I made sure it was most prominent in the calculation. But just typing in the UZR/150 ratings, you end up with real values like 17.2, -9.7, 0.5 (Cameron) for years that had a substantial sample size. Some kind of correction is needed for players with no or little data. Since someone intelligent chose them to play a substantial amount of time in CF, I thought it best to push them toward average. I'm open to suggestions.
Add them together. They are both stats that produce runs over or below a league average, over a season.