Thinking Way Outside the Box: Trader Jack and The D-Train

Florida
1. J. Pierre, CF
2. J. Conine, LF
3. C. Delgado, 1B
4. J. Encarnacion, RF
5. P. LoDuca, C
6. M. Lowell, 3B
7. D. Willis, P
8. J. Dillon, 2B
9. R. Andino, SS

That was the lineup posted on the wall of the visiting clubhouse at Shea Stadium prior to Thursday night’s game between the Florida Marlins and New York Mets. It did not cause the ground to open and swallow both teams. The polar ice caps did not melt. The Earth did not break free from its orbit and crash into the sun. Yet a fundamental tenet of baseball, as clear and without question as the law of gravity, had been brazenly violated. The pitcher–long since banished by the baseball gods to the furthest rung of the batting order–had been elevated to the seventh spot in the lineup.

How could manager Jack McKeon attempt such a stunt in the middle of the pennant race, with the Marlins, four back of the Houston Astros in the wild-card hunt, fighting for their playoff lives? The pitcher has always hit ninth, quite obviously because he is almost always the worst batter in the lineup. In the sometimes complex theory of lineup construction, this concept seems a no-brainer: hit your weakest hitter last, so he receives the fewest plate appearances, contributing as little drag as possible on offensive production. Moving him one spot higher in the lineup leads to roughly 20 more plate appearances a season, so bumping the pitcher up is a needless expenditure of outs, the most precious of the game’s resources.

Dontrelle Willis, however, is not a typical pitcher. He’s above average in multiple respects–on the hill (third in the majors in Support Neutral Value Added, 5.1 wins better than the average starter) and also at the plate. After Thursday’s 1-for-4 performance, Willis is 22-for-88 on the season with one homer and four doubles, good for a batting VORP of 9.6, better than all pitchers except Jason Marquis (16.4) and Carlos Zambrano (13.5). Willis’ home run came against Houston on September 12, when he threw a complete game and hit a two-run shot in an 8-2 victory. That dominating performance prompted Trader Jack to move Willis into the eighth spot of the lineup on September 17 against Philadelphia, ahead of shortstop Robert Andino, the first time a pitcher hit eighth since May 30 of last year, when Montreal’s Tomokazu Ohka batted in front of second baseman Jamey Carroll (a head-scratcher from manager Frank Robinson, considering Ohka’s career 312 OPS).

Since 1972, the first year for which Baseball Prospectus has play-by-play data, the pitcher has hit eighth 81 times. Seventy-seven of those came in the 1998 season, when Tony LaRussa irritated the baseball world by slotting his pitchers ahead of catching tandem Eli Marrero and Tom Pagnozzi in an effort to get more men on base for Mark McGwire during McGwire’s otherworldly season. The strategy had mixed results; although the Cardinals bumped their run scoring from 4.9 per game before the experiment began on July 9 to five per game with the pitcher hitting eighth, McGwire’s RBI/game fell from 1.09 to 0.80.

While this looks like LaRussa’s trademark over-managing, he may have been on to something. James Click’s recent article on lineup construction suggests that a slight increase in scoring can be expected when the pitcher bats eighth (although in James’s study, it only increased the mean by two runs), due to stringing a more potent on-base threat right before the better run producers at the top of an order.

So there’s precedent for batting a pitcher eighth, even a modicum of statistical justification for it. But seventh? To be fair, extenuating circumstances helped to force McKeon’s hand. With regular shortstop Alex Gonzalez and second baseman Luis Castillo sidelined, the skipper had experimented with Mike Lowell at second base in Wednesday’s game. That plan was scuttled when third baseman Miguel Cabrera left with a knee injury, forcing McKeon to trot out two players with little MLB experience–Andino and Joe Dillon–in the middle infield.

The last time a pitcher hit seventh or higher was in 1973, in an August 26 game between the Padres and Expos at Parc Jarry. Expos’ pitcher Steve Renko batted ahead of second baseman Pepe Frias and catcher Terry Humphrey. Renko finished the year hitting .273/.333/.341, which for the ’70s would have been a highly serviceable line from a middle infielder. The move was justified–Frias hit .231/.266/.284 in ’73, his rookie year, and Humphrey hit .167/.206/.222. Renko went 2-for-3 with a double, while Frias and Humphrey combined to go 2-for-7.

It’s tough to distinguish Andino and Dillon from their Padres counterparts, as the two Marlins have combined to come up with just nine hits in 50 at-bats on the year (the pair went 1-for-6 on Thursday). Andino, a 21-year-old September call-up with a strong defensive reputation, likely deserves to hit ninth when Willis takes the mound. He carried a career minor-league line of .236/.290/.315 in 970 at-bats in the low minors into 2005, while Willis owns a line of .232/.265/.318 in 220 AB in the majors. Dillon is a different story. The 29-year-old veteran of eight seasons in the bush has a career .514 minor-league slugging percentage and 131 homers. Out of baseball and coaching at Texas Tech in 2003, Dillon has fought all the way to the majors after a monster 2004, only to suffer the indignity of having to hit behind a pitcher.

As long as the Marlins are shorthanded in the infield, eighth seems to be the spot for Willis, both because of his strong bat and the small statistical advantage of what James Click termed “LaRussa’s Gambit.” After first batting Willis eighth last Saturday, McKeon exclaimed “If he wasn’t a pitcher and you gave him a little time, he could be a first baseman. He could hit in the big leagues. He doesn’t miss many balls when he swings the bat.” Holding down first base in the majors is likely beyond even Willis’s bountiful skills, but McKeon’s willingness to loosen up on the reigns of convention, fearlessly trying a controversial strategy even in the midst of his team’s playoff run, should compel other managers to look with a more calculating eye at lineup construction.

James Click and Jonah Keri contributed research to this article.

Caleb Peiffer is a contributor to Baseball Prospectus. You can reach Caleb by clicking here or click here to see Caleb’s other articles.