NUMBERS GAME

Evaluation By Numbers Is Beginning To Add Up (Thomas Boswell, May 29, 2003, The Washington Post)

Guess what: "Revenge of the Nerds" may be playing in a ballpark near you. [...]

* Analyze all hitters through on-base percentage. Getting on base, while making the fewest outs, is the heart of offense. Walks are wonderful. Hitters who know the strike zone drive pitchers crazy. High on-base hitters usually take many pitches, foul off two-strike pitches and, as a result, exhaust the pitch limits of most quality starters. Result: Crummy relievers enter the game and get waxed. Even in a three-game series, the high on-base team wins a war of pitching attrition. The Yankees teams of Joe Torre have used this theory in recent years. The Red Sox do now.

* Slugging percentage is the only other vital offensive statistic. Power matters. Combine on-base and slugging averages, with much more emphasis on the former, and you'll automatically build a high-scoring lineup. Hard as it is to believe, many high on-base average players come cheap. Walks are boring. Nobody comes to see, or pays big salaries to, walkers. (So, Beane grabs 'em easily.) As for batting average, ignore it. Irrelevant. Forget stolen bases, too. Until your success rate is over 70 percent, attempting to steal is, mathematically speaking, a waste of time.

* A superstar, such as Giambi, can be replaced -- at reasonable cost -- in pieces. When the A's lost Jason Giambi, Johnny Damon and DH Olmedo Saenz after '01, they added David Justice, Hatteberg and Jeremy Giambi. The combined on-base percentage and slugging percentage of the three new players roughly equaled the comparable statistics of the three lost players. Jason Giambi wasn't missed.

* Any decent pitcher can be turned into a star closer because any solid pitcher should be able to pitch one inning when he always enters with the bases empty. Once you create such an overrated star, you immediately trade him at peak value. Then just develop a new closer since it's so easy to do. Repeat as needed.

* To evaluate pitchers, use the breakthrough DIPS theory of stat man Voros McCracken that's been invented in the last three years. DIPS stands for "defense independent pitching statistic." It's a stunner. Nobody believed it at first, but now most serious stat geeks accept it. Once a batter hits a pitch, it's very close to pure luck whether it gets caught or not. From one season to another, for example, Greg Maddux's ERA may fluctuate by 1.5 runs even though he pitches identically. One season a lot of hits find holes. The next, they don't.

Like most baseball geeks, I've been a huge fan of Bill James and Tom Boswell for twenty years, and have long believed in their numbers, even when they suggest extraordinary things--like that Robbie Alomar, because he gets to so few balls, is a below average second baseman. So, as Mr. Boswell says, it's been a vindication this year to watch the Red Sox, under the insanely courageous young Theo Epstein, put these kinds of statistical analyses to work and have them work. Posted by Orrin Judd at May 29, 2003 1:10 PM