MLB Pace-of-Play Basics and Why the Rule Changes Can’t Work
Update: Charts use 2015 data through June 2.
A lot is being made of MLB’s new rules that are intended to solve the “problem” that games take too long. But there’s nothing MLB can do to significantly shorten the average game duration, and we all need to understand why.
It’s in the data.
The most rudimentary possible measure of the pace of play is the most often cited: average game duration. This is by far the easiest to capture and the easiest to understand, but it tells us almost nothing. It doesn’t account for any variation in game play, such as offensive production, baserunners, number of batters, pitches; it doesn’t bother to differentiate between pitchers’ duels and slugfests, assuming that all of these things average out over the course of a season. They do, but this doesn’t mean the game hasn’t changed over time: it has.
Baseball Prospectus has average game duration stats that go back to 1950. (I can’t vouch for any of the data cited here, and there are some data points that make me go hmmm, but I don’t have time to verify it either.) Here’s what the game duration data looks like:
I’d throw out the 1950s data because it’s so choppy, but from 1960 it looks pretty good. Looks like an increase from about 2:35 in the 1960s to about 3:00 today.
But like I said, this is the most rudimentary data possible. What changes in the game, other than extended crotch-scratching, can explain games taking so much longer?
The most basic unit of a baseball game is the pitch. We don’t have per-game pitch data back to the 1960s, but we do have it since 1988, when STATS, Inc. started collecting pitch data from every game. Here’s the number of pitches per game since 1988:
For the 10 years starting in 1988, games averaged 269.5 pitches. But from 1997 to 1999, games jumped from 268 to 292.5 pitches, and have never since fallen below 286 pitches per game over any season. The number of pitches per game shows a clear increase over time.
Now put the two data sets together, dividing average game length by the number of pitches per game to get seconds per pitch (sec/pitch), which is a much better measure of pace of play than game length alone:
Sec/pitch generally decreased from the late 1980s to the early 2000s, then slowly increased. So while game length is significantly longer in the last few years compared to 25 years ago, compare sec/pitch during the 2012-2015 period to the 1988-1997 period:
While game length has increased, the number of pitches per game has increased more, and time per pitch has decreased. Since 2012 seconds per pitch has averaged 37.94, which is faster than seven of the 10 years starting in 1988.
Why does this matter? Because it shows that pace of play is not the problem it’s made out to be.
Games are longer today because the game is played differently today. Games are longer because players and managers use data that wasn’t available decades ago to make smarter decisions (in general), which means the game is played better now than it was played in the past. There are many more pitching changes today. Batters see more pitches. Maybe the strike zone has changed and pitchers nibble more for the unhittable off-the-plate areas that are now routinely called strikes. There are lots of reasons games take 25 minutes more today than in the early 1960s, and almost none of them can be changed by MLB’s pace of play rules–nor should they be.
Major League Baseball can’t significantly shorten games without fundamentally affecting the way the game is played. This doesn’t mean they shouldn’t try their little shot clocks and keep-one-foot-in-the-batter’s-box rules; it just means we shouldn’t expect much in the way of results.