For all of his conservative ways, Tony La Russa has also been an innovator. He has institutionalized the LOOGy, recast career relief pitchers into starters and, most recently, begun a trend (well, a trend of one) of batting the pitcher eighth in his lineup. While not all of his decisions are sound (see Cesar Izturis, leadoff man) and his rationale isn’t always spot-on, La Russa has at least shown that he can try something new occasionally.

We noted recently that one of the Cardinals’ competitive advantages is that they have a relative hoarde of middling starting pitchers, half of whom are due to return from injury at various points this season in a kind of planned continual rollout. That gives La Runcan the following options:

1. Pick the best five starters and trade one or more of the others.
2. Pick the best five starters and move one or more of the others to the bullpen.
3. Optimize on the best set of pitchers in some other way.

We’ve championed the radical idea of a 10-man rotation in the past, though it may more flexibly be called a "six-plus rotation." At any rate, we thought we’d attempt to quantify some of the advantages. Today, we’ll look at the proposed gain in offense from reducing pitcher plate appearances. One of the theoretical benefits of our plan is that a team will realize a gain in offense because, by changing pitchers earlier in the game — such as before the pitcher faces a batter more than once (or, at most, twice) — the team’s pinch hitters will produce more offense than the pitchers would have. So is that true?

Over the last five years, NL pitchers have averaged about 2.3 plate appearances per game. Those at-bats aren’t automatic outs, but they’re pretty close to it, given that the pitchers offer around a .126 Gross Production Average (NL average to-date this season is .251. Using GPA, we’ve calculated the Runs total using the formula R=PA*1.356*(GPA^1.77):

Over the last five seasons, then, the average yearly amount of runs that pitchers as batters contributed to their teams (batting around 2.3 times per game) was 13, or about 1.3 wins.

So could a team improve upon that 13 runs/season by having its pitchers bat less often? Granted, 2.3 plate appearances per game doesn’t sound like much, but over the course of a season it translates to around 377 plate appearances per team. What if a team decided to simply reduce the number of times per game that its pitcher bats, say, to once per game? The other roughly 1.3 plate appearances would be taken by a pinch hitter. Assuming that pinch hitters are better hitters than pitchers, how many runs would such a strategy gain the team?

We’ll run the same numbers from 2003-2007, reducing the pitcher’s plate appearances to 1.0 and giving the remainder 1.3 to pinch hitters, whose actual OBP/SLG/GPA numbers are used:

So splitting the normal pitcher plate appearances between the pitcher and pinch hitters, teams would’ve realized a total on average of 26.5 runs over the season, or 13.5 more runs than they actually did by giving those at-bats only to pitchers. Converting to wins, that’s an extra 1.3 wins — essentially for "free."

Implementing such a strategy would of course curtail the number of innings a starting pitcher (or any pitcher, for that matter) could go. Assuming that the pitcher bats eighth or ninth, the starting pitcher could at most pitch five innings. Taking into consideration the fact that hitters are increasingly lethal in their second and third at-bats against the same pitcher, that’s not exactly a bad thing.It would, however, force teams to rethink their bullpens; relievers would need to be able to throw more than one inning at a time, for instance.

But say that a team did try the strategy of limiting pitcher at-bats. Once the team got a feel for the right roster composition, might it not take the next natural step? That is, why not go all the way and eliminate pitcher plate appearances altogether? If teams could realize an extra win and a half from limiting the pitcher to one at-bat per game, let’s see how much they could benefit from pinch-hitting and/or double-switching every time the pitcher’s spot came up; we’ll substitute the pinch-hitting stats for all of those 2.3 PA/G:

Over the last five years, such a strategy would’ve resulted in a net 23.7 runs per season, or about two and a half wins.

So whether a team decides to eliminate the pitcher’s at-bats altogether or merely reduce them to one per game, that team stands to gain between 1.3 and 2.4 wins per year. That’s not chump change. Considering the fact that the Cardinals (and now the Brewers) have implemented a somewhat radical strategy that promises less — batting a pitcher eighth can yield around 15 extra runs per season, or 1.5 wins — it seems like an easy decision. We’re not going to hold our breath that the Cardinals (or anyone) will cop to it anytime soon, but it certainly makes the six-plus rotation more appealing from an objective, quantifiable perspective.

This entry was posted by Pip on Tuesday, April 15th, 2008 at 2:47 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.