Printer-friendly

February 14, 2008

Schrodinger's Bat

Wrapping it Up

by Dan Fox

"Baseball is Darwinian in its results but Newtonian in its processes. With rigorous empiricism fielders built up a solid foundation of knowledge through generations of trials and thousands of errors. Their peer-reviewed results quickly became standard operating procedures."
--Cait Murphy, describing the evolution of defensive play in Crazy '08: How a Cast of Cranks, Rogues, Boneheads, and Magnates Created the Greatest Year in Baseball History

---

Over the last couple of months, we've been playing around with measuring the efficacy of those "standard operating procedures" by using Retrosheet-style play by play codes, i.e. rather coarse records of each play, including hit type (line drive, groundball, fly ball, or popup) and the fielder who picked up the ball, but not batted ball location information. Because this method doesn't use the most fine-grained data in existence, I've dubbed the system Simple Fielding Runs, or SFR for short. The system for infielders has gone through a couple of major revisions that have been fully discussed here, while the system for outfielders has also been revised from the ground up.

This week, we won't be delving deeply into the methodology and making tweaks at the lowest levels of the system. After the results from last week, while we might still consider just why right fielders like Brian Giles and Juan Encarnacion show such large differences between SFR and UZR (but thank you to the readers who offered observational evidence of both), I'm pretty satisfied that the system works well enough to be useful. So this week I haven't made any significant changes to the code.

Of course, at this point I've calculated data from five seasons (2003-2007) for major leaguers, and since the outfield system is heavily dependent on the inherent park effects that are captured by comparing how a particular fielder performed in a particular park to how all other fielders did (inspired by an approach from a reputable source), adding more data in the future will tend to smooth out the results. That said, I do want to break down the components of the outfield SFR system a little, look at a rate statistic, evaluate team totals, and combine SFR with throwing before putting a bow on top of this confection and finally giving you all the numbers.