ST. LOUIS - Terry Collins leaned forward in his chair earlier this week and peered through the reading glasses that had slipped halfway down his nose. He scanned a color-coded sheet of paper -- red for hot hitters, blue for cold -- no different from the hundreds of pages of statistics that he uses to shape his lineup.
A team stocked with productive hitters can absolve a manager of this tiresome exercise. But this has become a daily ritual for the Mets manager -- even though it ultimately may not make much of a difference.
"It should take me five minutes to write it," Collins said Tuesday. "It took me an hour and 15 today."
Not including the pitcher, the Mets used 63 starting lineups in their first 70 games, though they continue to endure dry spells at the plate. The latest wrinkle came Monday night when Collins batted righthander Jacob deGrom in the eighth spot. It was the first time in franchise history -- 8,354 games -- that the Mets batted their starting pitcher anywhere else but last.
Tuesday night, Collins went to the same well again, batting Jonathon Niese eighth.
"All I know is that there's a guy in the Hall of Fame who did it, so it can't be all bad," Collins said, a reference to Tony La Russa, who employed the tactic while managing St. Louis.
There is evidence to suggest that moving the pitcher into the eighth spot achieves a double-leadoff effect, with the No. 9 hitter functioning as a table setter for the more dangerous bats at the top and middle of a lineup. But statistics also show that lineup construction overall has relatively little impact on run scoring.
If anything, the numbers reinforce what always has been accepted as common knowledge: The order of the bats is secondary to the bats themselves.
"I can't put in different names," Collins said. "The names are all the same."
Part of the reason is that lineup decisions often aren't decisions at all.
In theory, a manager has a wide range of possible combinations with nine spots to fill. In reality, the number of choices is much smaller. For instance, no big-league manager would ever consistently bat a star player in the ninth spot, nor would he have the pitcher hit cleanup.
Once extreme lineup constructions are taken out of the equation, what's left is a relatively small range of choices. And among these combinations, research has shown that the difference between the best and worst possible lineups account for no more than a handful of runs. And that's over the course of an entire season.
For example, according to the lineup analysis tool created by statistician David Pinto, the Mets' lineup on Monday with deGrom hitting eighth was projected to score 4.649 runs per game (an artificially high number since deGrom's on-base percentage (.462) and slugging (.500) are inflated in a small sample). With deGrom hitting ninth, the projection was 4.621 runs.
The difference over a full season? Five runs.
If there is a gain to be made in lineup construction, it's based more on getting the best hitters more plate appearances, and less on other elements such as speed or contact ability. Though Eric Young Jr. has elite speed, his .313 on-base percentage makes him a bad fit as a leadoff man.
Nevertheless, the lineup carousel likely will roll on. Though it may not make a difference in the long term -- an argument that Collins isn't completely sold on -- his focus remains on the impact for that day. It's why he arrived at the ballpark at 11 a.m. and tinkered with the lineup until 12:15 p.m.
Said Collins: "We've got to try something."