February 8, 2010

Baseball Therapy

Why Not Two Pitchers?

by Russell A. Carleton

Printer-friendly

Contact Author

It was the second game of a doubleheader last July 12, and the Cardinals were visiting Wrigley Field. In the top of the ninth inning, the Cards held a 4-2 lead, and the wheels were moving in the head of Cubs manager Lou Piniella. Piniella had brought lefty Sean Marshall into the game with runners on first and second and no one out to face the announced left-handed hitting Cardinals pinch hitter Chris Duncan (Tony La Russa countered by using Nick Stavinoha to pinch-hit). Marshall walked Stavinoha, and Piniella popped out of the dugout and called to his bullpen. In came the right-handed Aaron Heilman to face Brendan Ryan, and Marshall was dismissed from the mound to… left field. It was Cubs left fielder Alfonso Soriano who was headed for the showers, rather than Marshall. Piniella apparently wanted to keep Marshall in the game to face Skip Schumaker and Colby Rasmus, the next two hitters due up after Ryan. Piniella's strategy worked. Heilman struck Ryan out. Marshall then returned to the mound. He ended up striking out Jaret Hoffpauir (pinch hitting for Schumaker) and getting Rasmus to fly to left field, where Marshall's replacement, Reed Johnson, made a fine catch.

I happened to catch the telecast of this game flipping through the channels, and after it happened, I wondered why this doesn't happen more often. I knew of one other example from dithering around Retrosheet (yes, that's what I do in my spare time). On July 22, 1986, the Mets were involved in an extra-inning game with the Reds, and due to a few well-placed ejections after a brawl, found themselves short-handed with only seven position players in the game and no one left on the bench. Their solution: bring in Jesse Orosco and Roger McDowell and have them take turns pitching and playing in the outfield. At one point, McDowell switched back and forth between left and right field with Mookie Wilson, presumably so that he was in the batter's opposite field and was less likely to be tested. As such, the box score from that day makes for a rather interesting read. The Mets had the final laugh in the end and won the game in 14 innings. And the 1986 World Series.

I've often wondered why this sort of thing doesn't happen more often, particularly in the National League. There have been other instances, although relatively few. I'm not thinking that it would be an everyday occurrence, but it shouldn't be so exotic that I should write an article about it, either. It would require a particular set of circumstances, but one that would probably come up a few times each year. And it would probably benefit the team that used it, perhaps a great deal.

I propose that teams should use a twin-reliever strategy more often. It may seem a little odd, but it fits within the rules. A team would have a right-handed reliever and a left-handed reliever in the game at the same time, and one would play in the field while the other pitched. They could switch as the situation dictates.

First off, let's consider the mechanics. Left and right field make sense as a hiding place for someone whom you suspect will be a poor fielder. Left field in particular is the type of position where teams put guys when they have no idea what else to do. Chris Coghlan, last year's NL Rookie of the Year, had played all of three games in the outfield in Triple-A. A natural second baseman, his path to the majors was blocked by Dan Uggla. So, the Marlins shrugged and sent him to the outfield where he got on-the-job training. And he did fine.

(A side note: In dire circumstances, when position players are called on to pitch, it's often a player who "pitched back in college/high school/little league." This is considered appropriate justification for such a move… OK, that and the fact that it's the 23^rd inning. But many pitchers played in the field during their school days. Perhaps even in the outfield. Why should it be a shock that they might take a few reps out there now?)

Putting this strategy into effect means that a team has a left fielder who they don't mind pulling out of the game. If he's a spare part anyway, he probably won't be missed. Alternately, the spare part guy could be at another position, and the left fielder might shift there. The team will eventually need another "real" left fielder (or maybe not!) who will likely be a replacement-level player (by definition). But if a team is already living with a starting left fielder who is basically a replacement-level player, then the team has nothing to lose by taking him out of the game.

The strategy would probably also work best for a team that has only one lefty in the pen, or perhaps only one lefty that they trust. In a game against a team with a bunch of tough left-handed hitters, especially those who are staggered throughout the lineup (i.e., the lineup goes left-right-left-right-left-right), a manager is faced with the choice of putting in a new pitcher for every batter or letting his LOOGY face a few righties (and throw a few extra pitches) as well.

Finally, our twin-reliever strategy probably should also be kept for times when the pitcher's spot in the batting order has turned over. If the left fielder is batting eighth, that makes it all the better. This way, both relievers can be kept in for two innings or so before mangers have to start thinking about pinch hitters. If the team employing the strategy goes crazy and sends eight guys to the plate in the top of the next inning, it means a few runs have scored and it's probably time for a new bullpen strategy anyway, and for what it's worth, the manager can keep one of the relievers in the game if he chooses. The strategy would not work as well in the American League. If the pitcher moves to a defensive position, his team loses the designated hitter. While this strategy has been used in the post-DH AL (by none other than Piniella with the 1993 Mariners), it is likely to be a creature of the NL. Perhaps an AL team might use it in the ninth inning.

Here's how it would work. The first reliever would enter the game as he normally would and takes the pitcher's spot in the order (ninth, presumably). When it comes time for him to yield to the second reliever (of the opposite handedness), the first reliever trots out to left field, while the second reliever enters the game in the left fielder's batting spot (eighth?). If they go out for a second inning, one can take his normal warm-up tosses from the mound, while the bullpen catcher goes out and "plays catch" with the "left fielder" … in a crouch at a distance of 60 feet.

What's to gain? The obvious reason is that a team could have the platoon advantage against every batter (except switch hitters, but they could pick which side he would bat from). In 2009, left-handed hitters had an OBP of .317 against left-handed pitchers, but .347 against righties. Right-handed hitters showed a similar pattern, (.318 vs. .341). So, the platoon advantage is worth roughly .025 outs per each extra batter who gets "platooned." How much would having a non-professional left fielder cost? I created a toy defensive system called OPA! (out probability added above average) a few years ago. In 2009, the worst left fielder in baseball was Carlos Lee. He cost the Astros about .017 outs per inning last year with his fielding compared to the average left fielder. Even if our moonlighting pitcher is equal to the worst left fielder in baseball (or a little worse), the twin-reliever system more than pays for itself in one extra batter for which you can gain the platoon advantage (assuming that the replaced left fielder was average, of course).

There are also a few indirect effects. One is that it saves other members of the bullpen from having to make an appearance. In the event of an extra-inning game, the team has not burned through its available pitchers by using four relievers in the eighth inning alone. The LOOGY can face only lefties, and concentrate all of his allotted pitches on them, and the manager can stick with a righty that he trusts rather than having to bring in two or three of them. The one disadvantage is that it shortens a team's bench. You'd have to pinch hit for the pitcher's spot anyway, but now you have two non-hitters to think about. Then again, if a team did it at the end of a game (say, the eighth or ninth inning), it might not be an issue.

The math here isn't that hard to fathom, and surely someone has thought of this along the way. The question is why it doesn't happen more often. Again, I don't suppose that it would happen regularly, but neither does a suicide squeeze, and everyone knows what that is. It is something of an unorthodox strategy, and in baseball, those are (oddly) often looked upon with a little bit of suspicion. It's not that the rewards of such a strategy would be huge, but they would be net-positive. Perhaps teams are unwilling to change their game plan for what is likely a modest gain in win expectancy. However, the only cost to the team is a little more thought put into actually putting the strategy into effect. A team has very little to lose and something tangible to gain. Why wouldn't a manager do something clearly in his team's best interest?

Maybe it has something to do with the fact that this is not what people are used to seeing, and it takes intestinal fortitude to introduce something new. There will be nights when the twin relievers fail to hold a lead, and people will (illogically) assume that it was because of the use of the twin-reliever system. The manager will be decried as someone who "tinkers too much" and who should "stick to managing by the book" by the 40,000 other managers who paid to get into the stadium. One must be called a fool before being called a genius, and few people can handle being called a fool. But here's a chance for a manager to make a mark as the guy who popularized this strategy, and maybe won an extra game or two for his team.

Russell A. Carleton is an author of Baseball Prospectus. Follow @pizzacutter4
Click here to see Russell's other articles. You can contact Russell by clicking here