Statistics and Baseball
by Mischa Gelman
(published 10/11/2000)

    In the past 20 years, baseball analysis has seen an increasing usage
of sabermetrics, the application of statistics to the American pasttime.
This change has even begun to lead to different administration of certain
teams - several have hired sabermetrics as consultants and advisers, and
one general manager, Billy Beane of Oakland, has been a strong advocate of
the sabermetric approach to the game. Do the statgeeks have it right? Or
do they overlook elements that cannot be quantified?
    Oakland is a great team to look at to see how well the methodology
works, given their wholesale support of the approach. Both manager Art
Howe and Beane have bought into the notion that OPS (on-base percentage
plus slugging percentage) is the simplest guide to offensive ability and
to the idea that pitch counts are essential for young pitchers. The A's
have certainly made their strides the past four years - given up for dead
in the mid-90s, they bottomed out at 65-97 in Howe's second season, 1997.
The team boasted such nobodies as Rafael Bournigal, Brian Lesher, Damon
Mashore and Mark Bellhorn. In '98, they improved to 74-88, as 22-year olds
Miguel Tejeda and Ben Grieve cracked the lineup (an optimistic faith in
youth also is obvious in many statheads). In '99, they upped their record
to 87-75, with 21-year old Eric Chavez and 23-year old Tim Hudson joining
the mix, plus stathead favorites like Tony Phillips, John Jaha and Randy
Velarde. Now in 2000, they continue the path upwards, going 81-68 and in
the heart of the wild card race as of this writing. Obviously, they have
been winning a lot more using the stathead approach. Walk-happy players
like Jaha, Velarde, Jason Giambi and Matt Stairs have helped the team be
productive. A guy like Stairs or Jeremy Giambi, who combines lots of walks
and homers with a low average, can be valuable, even if other teams fail
to recognize the fact. Hudson has been excellent at times, perhaps
indicating that pitch counts work - and perhaps indicating that he's just
a good pitcher.
    On the other hand, they have been far from dominant. The White Sox
have a better record despite low-OPS types like Chris Singleton and Mark
Johnson. The Yankees, predicted for failure this season by statheads,
remain atop their division, no matter how much they were derided for
having players who were too old. And the Astros, the statgeek-supported
team in the National League, has sucked, to put things succinctly.
    The sabermetric approach does seem to work in many ways - but it has
its flaws. Most noticeably, it overlooks elements like hustle and heart,
characteristics that made mediocre players like Mookie Wilson and Wally
Backman into valuable players to have on a team. The 1960 Pirates got by
on guts, not with overwhelming ability. The best team on paper doesn't
always win. The game, after all, is played on grass or artificial turf,
not paper. The approach of a Rob Neyer seems to be "well, I can't see it
or quantify it, so it mustn't exist." This reminds me of the atheist logic
that just because they can't see G-d, therefore G-d cannot exist. No
wonder all-out players like Jason Kendall (a statgeek favorite) are
critical of "little Rotisserie League geeks. They don't have the slightest
idea how to play the game except on a computer -- and it makes me sick." 
    Also, the sabermetricians are surprisingly blind to statistics that
don't correlate with their preconceived notions. There is no evidence
that high pitch counts affect those over 25, yet they still scream and
holler when a Kris Benson tops 100 pitches. For people relying upon math,
the usage of a round "magical" number like 100 is rather silly, when this
number is not supported by any research. Similarly, the relevation that
walks are underrated leads to SDCNs (stat-drunk computer nerds) to
overrate guys like Rob Deer who had low averages but walked a lot and
underrate high-contact, low-walk players like Kirby Puckett. They don't
seem to realize that Shawon Dunston was a good player in his prime,
despite drawing a walk once in a blue moon. In their pro-youth frenzy, 
they back minor-leaguers with mediocre track records (see Emil Brown, Adam
Hyzdu and Lou Collier for local examples) over proven major-leaguers like
Ed Sprague and John Vander Wal. For all their talk of impartiality, they
are just as much into mindless promotion of a given player as any other
fan (not that this is bad). And due to all of this ignoring of stats, and
due to their ignoring things they can't measure, their methods are limited
- they will lead to more wins, but they won't guarantee a darned thing.

   Mischa Gelman likes to play with numbers and likes to play baseball
(though he's much better at the former). He just doesn't think it's right
to worship one without looking at the other.