In my article Friday about what’s slowing down Major League Baseball games, I use a relatively new stat called “pace.” It’s a measure of how much time elapses between pitches in the same plate appearance, making use of the PITCHf/x tracking system installed in every major league park. It can be applied to pitchers for all batters they face; and to batters for their plate appearances.
One thing we already knew about pace is that it’s remarkably consistent from year to year. FanGraphs makes it possible to test that with a tool for calculating correlation between the same statistic across seasons. Pace’s year-to-year correlation is 0.859 among the 1,554 pairs of consecutive pitcher seasons for which FanGraphs has pace stats — from 2007 to 2013 for pitchers with at least 40 innings each season. That’s far higher than strikeout-to-walk ratio, which has a year-to-year correlation of 0.528; Fielding Independent Pitching (FIP), an ERA-like stat (0.443); winning percentage (0.081); and other fielding stats. The correlation persists over a gap of more than one season: It’s 0.757 from one season to another three years later.
That suggests that the biggest factor affecting pace is the pitcher. Pitchers can change catchers, teams or leagues; face a mix of batters; and pitch in front of a different defensive alignment or in different contexts. Yet their pace of play stays largely the same from season to season.
This is evident in the leaderboard: Five of the 10 slowest pitchers with at least 100 innings last year were among the 10 slowest in 2012.
Like other new stats, pace is a work in progress. For example, there are two versions — one on FanGraphs, the other on Baseball Prospectus — akin to the competing versions of wins above replacement. FanGraphs registers more average time, typically, which BP’s Ben Lindbergh attributes to his site’s exclusion of pickoff attempts.
One mystery is whether pace is associated with other pitching attributes. In a quick check using the FanGraphs correlation tool, it looked like pace was modestly correlated with FIP and Win Probability Added. But that was probably because relievers tend to have better stats than starters, and also a slower pace. There’s no reason to think one causes the other. When I restricted the sample to only pitchers with at least 150 innings pitched — thereby excluding most relievers — I found there was essentially no correlation between pace and FIP or WPA. Pitchers don’t seem to get any particular value out of a slow pace, though it’s possible they would have worse results if they had to hurry between pitches. In the spirit of their deliberate play, we won’t rush to judgment.
