Covering the Front and Back Pages of the Newspaper
March 3, 2003
BASEBALL: BP 2003
I recently picked up my copy of Baseball Prospectus 2003. A few thoughts:
1. The book now has actual stats, not just context-translated stats. I attribute this nod to conventional baseball-book practice mainly to the desire to snag more market share among the lucrative Rotisserie Baseball player market, as well as to the opportunity presented when the popular STATS, Inc. Handbook was pulled from the market by its new publisher.
2. No EqAs! EqA may well be a well-thought-out system, but I always found the translation into a batting average-type number distracting; I prefer measuring sticks like Runs Created/27 outs or Offensive Winning Percentage that enable you to explain the stat's thrust to a non-stathead without sounding like a monumental dork ("see, it's this measurement of offensive value pegged to a batting-average style scale . . . ") As with a few of the missing metrics, like the pitcher W-L approximations, I wonder if part of the idea was to make more things available only on the $39.95 subscription-only website.
3. No Pitcher Abuse Points. Personally I found the PAP system somewhat useful, but it was always obvious that the system was just a warm-up to a more systematic study. I suspect the PAP system was killed off by Will Carroll; I haven't digested much of Carroll's work yet, but he seems to be taking seriously a systematic study of injuries, a fruitful area for inquiry.
4. I like the new PECOTA projection system (I assume the acronym is a homage to Bill Pecota). As with most systems of this type, I don't have the technical expertise to judge the system itself, but as for its governing assumptions, I like the move to advance Bill James' Similarity Scores by looking at a player's prior 3 years rather than just his cumulative career numbers (this is a particularly significant advance for players over 30), as well as the system's recognition that it's projecting a range of possible outcomes and telling you the odds of each, rather than just making a single "projection." (It also seems that there are, for hitters, fewer stunningly optimistic projections, which especially in the case of young players has been a problem for past Prospectuses; you had to discount rookie projections if you were relying on BP for a Roto draft or risk getting roasted).
One minor bone to pick: Nate Silver identifies Similarity Scores as having been introduced by Bill James in The Politics of Glory, his Hall of Fame book published in the mid-1990s, when in fact -- although I couldn't immediately locate the reference -- James introduced them in one of his Abstracts back in the 1980s.
5. The snider-than-thou random potshots on non-baseball topics still appear throughout the book, and have started to wear real thin on me; it's not really a good idea to heap gratuitous insults on subjects that are not at hand, in my opinion, since you never know what your audience thinks about other topics, and why limit your baseball audience? (I know James did this stuff too, but he at least always had an air of the cranky professor giving up his personal biases and going off on long tangents, rather than just tossing out anonymous hit-n-run one-liners). On the other hand, the book seems almost conciliatory in dealing with baseball management -- the comment on Steve Phillips is a good example -- which could suggest a new humility, or could suggest that these guys are more 'insiders' now, more dependent on (and interesting in becoming) baseball insiders, and less willing to be nasty to them.