Tuesday, March 22, 2005
SLOBs at Play
by Tom Bozzo
One thing that crossed the table was the meaning of on-base plus slugging, or OPS, a baseball offense statistic that is defined as the sum of a player's "slugging average" and "on-base percentage." I like the Wall Street Journal's "SLOB" arrangement of the terms a bit better. Anyway, the question was whether OPS is totally ad-hoc.
Wikipedia, if you follow the link, provides the formula that results if you take a common denominator of the components and collect terms. The resulting fraction:
looks plenty ad-hoc.
If you redefine the terms somewhat, though, it does make more sense. Consider a modified slugging statistic SLG2 = (total bases on hits) / (plate appearances), which shares a common denominator with ob-base percentage.
That lets me write:
SLOB2 = SLG2 + on-base percentage = (2*H + Extra bases on hits + BB + HBP) / PA.
SLOB2 is just an offensive statistics index with "extra" weight on hits. It's obvious enough that reaching base on a hit may be more useful than reaching base on a walk — consider the potential effect on other base runners &mdash but less than obvious that SLOB2's parameters are the optimal choice for the more general :
SLOB3 = (a*H + b*(Extra bases on hits) + c*BB + d*HBP) / PA.
Some SABRmetrician has probably worked that one out.
Update: Edited a bit for clarity and to add links. Oscar has a fine baseball post up, too.
I'm sure you've all been waiting anxiously to hear about last night's Madison male blogger confab, with Freese and Oscar. Well, what happens at the Harmony Bar, stays at the Harmony Bar. Mostly.
One thing that crossed the table was the meaning of on-base plus slugging, or OPS, a baseball offense statistic that is defined as the sum of a player's "slugging average" and "on-base percentage." I like the Wall Street Journal's "SLOB" arrangement of the terms a bit better. Anyway, the question was whether OPS is totally ad-hoc.
Wikipedia, if you follow the link, provides the formula that results if you take a common denominator of the components and collect terms. The resulting fraction:
looks plenty ad-hoc.
If you redefine the terms somewhat, though, it does make more sense. Consider a modified slugging statistic SLG2 = (total bases on hits) / (plate appearances), which shares a common denominator with ob-base percentage.
That lets me write:
SLOB2 = SLG2 + on-base percentage = (2*H + Extra bases on hits + BB + HBP) / PA.
SLOB2 is just an offensive statistics index with "extra" weight on hits. It's obvious enough that reaching base on a hit may be more useful than reaching base on a walk — consider the potential effect on other base runners &mdash but less than obvious that SLOB2's parameters are the optimal choice for the more general :
SLOB3 = (a*H + b*(Extra bases on hits) + c*BB + d*HBP) / PA.
Some SABRmetrician has probably worked that one out.
Update: Edited a bit for clarity and to add links. Oscar has a fine baseball post up, too.
