Batting order is to a manager what wet clay is to a potter. It can take any number of shapes in the search for some elusive perfection. That's why Texas Rangers manager Buck Showalter keeps a notepad suction-cupped to his windshield. On the drive home after a game he starts scribbling ideas for the next day's work of art.
"People tell me, 'The traffic in New York must have been terrible,'" says Showalter, who managed the Yankees from 1992 through '95. "I say, 'I didn't mind at all. I got a lot of work done.'"
Casey Stengel, another former Yankees manager, would spread 12 lineup cards on the desk in front of him like a man playing solitaire. On each he would write the names of his two certainties-- Yogi Berra and Mickey Mantle--and then fill in various names in various places. After mulling his choices and imagining how they would play out, Stengel would pick the lineup that felt best.
More recently, after the Seattle Mariners had signed free agents Adrian Beltre and Richie Sexson this winter, manager Mike Hargrove sat down and wrote a prospective batting order. Then he tried another one. And then another. And another. This continued for a while. Hargrove estimates that he stopped after about 100 tries. "And you know which one I liked the best?" he says. "The first one."
Hargrove didn't come close to weighing all his options. Once a manager decides which nine players to start, he can pick from among 362,880 possible batting order combinations. If he were to consider playing each of the 25 players on his roster at each of the nine positions in the field, a manager would have 269,022,818,211,840,000 possible lineup cards. Even New York's traffic isn't that bad.
Ever since Alexander Cartwright decreed in his rules of baseball in 1845 that the players on each team would take turns batting in an immutable order, the batting order has been one of the game's great sources of angst for players and managers, and grist for sportswriters and fans--much of it unnecessarily so. For if the batting order is part science and part art, the scientific part tends to make the whole exercise look terribly overrated.
Mathematical studies over the past 40 years have found differences in reasonable batting orders to be negligible. The studies show that it is much more important whom you play than how you order them. Basically, as long as you don't do something stupid, such as batting your pitcher leadoff--or Washington Nationals shortstop Cristian Guzman second (box, below)--the difference between a lineup based on a manager's gut feeling and one that a supercomputer would statistically regard as the optimum has been found to be anywhere from two to 16 runs over a full season. That's it. (Statheads generally equate an extra 10 runs a season with one win.)
As if to prove the point (though in reality, he did so to break a slump), former manager Billy Martin more than once let his starting nine pick their spots in the batting order by drawing numbers out of a hat. He did so on Aug. 13, 1972, in the first game of a doubleheader, while trying to snap his Detroit Tigers' four-game losing streak. Detroit won 3-2 as cleanup hitter du jour Ed Brinkman, who hit 60 home runs in 6,045 career at bats, drove in the tying run and scored the game-winner. Martin returned to his conventional lineup in the nightcap, and the Tigers lost 9-2.
Two seasons later, when he was managing the Rangers, Martin let pitcher Ferguson Jenkins bat instead of using a designated hitter. Jenkins, one of the better-hitting pitchers of his day, broke up a no-hitter with a single in the sixth inning, scored the first run and won 2-1.
In 1964 a retired metallurgist named Earnshaw Cook espoused in his book Percentage Baseball that a manager should order his hitters from best to worst. He figured that such an arrangement would produce 11 more runs over a season.