The Rockies experiment with a four-man rotation was perhaps the strangest development during the disastrous 2012 season. Widely mocked, the four-man rotation with its 75 pitch limit did not last long and the Rockies are on the record as going forward this year with a traditional 5 man rotation with a 90-100 pitch limit. What has not received enough attention was the motivation behind making the change in the first place. Was the goal to limit injuries? Play to a strength of the team (the bullpen)? Perhaps the most compelling reason for the low pitch limit was to restrict the amount of times a starting pitcher must face the opposing team. As a pitcher progresses through the lineup multiple times, his performance tends to decline. This happens for any number of reasons including pitcher fatigue and batter experience: a slider at pitch 87 is going be more hittable than at pitch 11, especially when a batter is seeing it for a third or fourth time. This notion is also supported statistically:
The theory of increasing pitching efficiency by limiting lineup exposure was formally presented in March 2011 by Greg Rubin at the MIT Sloan Sports Analytics Conference. Rubin theorized that using a pair of starters each pitching 4 innings is more effective than a single ace. Rubin’s calculations led to the conclusion that a paired pitching strategy would reduce the runs allowed by a pitching staff by 96 runs over the course of the season. The four-man rotation from the Rockies was – intentional or not – a real world test of Rubin’s hypothesis. As we know, the physical demands from the Rockies’ implementation attempt were too great: the starters were unable to stay healthy with just 3 days between starts. However, the failure of the four-man rotation does not mean that the limiting exposure strategy is inherently flawed. This may seem obvious, but it is an important distinction to make: paired pitching or a four-man rotation are only ways to implement a limited exposure strategy. The overarching idea behind the strategy is bigger than any one attempt at implementation.