Way Too Much on Stats that Don’t Matter
June 4th, 2009 | Published in Daniel Gettinger, Padres, baseball, links, rants | 4 Comments
by Daniel Gettinger
In today’s UT is an article titled “Hairston’s Injury Leaves the Pads without Pop.” In the article, author Bill Center tries to demonstrate just how valuable Scott Hairston has been to the Padres this season. Center’s observation that “the only hitter more valuable to the Padres is Adrian Gonzalez,” is dead on. The manner in which he proves his case is not.
Center first attempts to demonstrate Hairston’s worth by pointing out that he has driven in the second most runs on the team (behind only Adrian Gonzalez). I don’t like using RBI’s when discussing the quality of a player’s season. It is a statistic that is too context dependent. But, if Center had moved on to more enlightening stats, I would not have minded the quick mention. Unfortunately, things only got worse.
Center’s next statistic of choice was batting average:
“Hairston carried the Padres’ top overall batting average (.327), the highest at Petco Park (.311) and the highest against left-handed pitchers (.379).The latter mark was 95 points over Gonzalez’s second-highest average against left-handers and 161 points above the team norm. And Hairston’s mark at Petco Park was 93 points above the team average.”
Umm…okay. These numbers are, all else equal, pretty good. But we know that batting average does not tell the complete story. Why not use on base percentage or OPS, both of which explain run production better than batting average, and are becoming mainstream statistics. It’s not like I am calling for the use of something like wOBA.
The final straw was Center’s decision to point out Hairston’s success in various small-sample size situations. He wrote:
Hairston was hitting .414 with runners in scoring position (12-for-29) with three homers, 17 RBI and a .793 slugging percentage and was 6-for-11 in RISP situations with two outs.
Overall, Hairston was a .395 hitter with two outs; .a 379 hitter in tie games, and had a .329 average with five homers and 16 RBI in Padres wins.
Did you also know that Hairston is batting only 0.250 (9-for-36) when he leads off an inning? Or that as a left-fielder, he is batting 0.217 (10-for-46)? Splits are fun things to look at, but with the exception of righty-lefty matchups, they don’t really tell us anything about the player. The differences are almost always a function of small sample sizes. Nothing more.
Scott Hairston has played well this season. I applaud Center for attempting to convey to UT readers just how good Hairston has been. I just wish he would have done so using statistics that properly capture the quality of Hairston’s play.
June 4th, 2009 at 2:57 pm (#)
I wonder how hip and new OPS really is these days. Lawd, they show a player’s OBP for every at-bat on TV!
Also, splits are funny things. I like using BA with RISP for when someone tries to claim the Hated Player of the Week isn’t “clutch”. Pitching with RISP is difficult and batters almost always fare better in those situations. It may not be an accurate way to show that someone is or isn’t clutch, but it quiets someone down who makes that argument.
But back to splits. I get the feeling people like using them to tell a story, or to agree with preconceived notions, not for honest player evaluation. If you’re going to unnecessarily limit the population size, there should be a good reason. In other words, evidence that the split needs to be split.
June 4th, 2009 at 4:02 pm (#)
“I wonder how hip and new OPS really is these days.”…My point exactly. Fans now seem very comfortable with both measures (Tom Krasovic used them extensively), and either one would have been a better choice than batting average.
Also, really good point when you say “if you’re going to unnecessarily limit the population size, there should be a good reason.”
June 4th, 2009 at 5:00 pm (#)
[...] Way Too Much on Stats that Don’t Matter :: Friar ForecastBill Center from the UT isn’t using the new hip stats. [...]
June 6th, 2009 at 1:48 pm (#)
Me thinks thous does complain to much ..
Have ya ever read the comments on a typical union tribune piece? Its Stephen King scary. Those are the people he is reaching. Those bozos are his target audience. Sorry it just is. I imagine most of them had to look up RISP, as the thought it was a new form of surfboard wax.
to them, all the numbers were a Rorschach test