The Futility Infielder

Falling Pitch Counts

The relationship between a pitcher's workload and his tendency toward injury is perhaps the most controversial area of inquiry among those who study baseball statistics. The studies haven't proven much, but lots of blood has been spilled among researchers over the matter. The argument has been rather impassioned, primarily because of what's at stake: protecting young pitching arms from overuse and injury.

The prevailing school of thought is that throwing beyond a certain threshold in a single game increases the risk of injury. A study published in Baseball Prospectus 2001 and cited by Don Malcolm in this Baseball Primer piece suggests that above 122 pitches, there is a "moderate risk". The link between correlation and causation seems to be the bone of contention. I'm in no position to summarize the arguments here, being a late-comer to the party. But the anecdotal evidence, especially with regards to young pitchers and repeated abuse, is compelling.

Pitcher workloads, based on number of pitches per game and innings pitched per season, have been in decline for a long, long time. The rise of the relief specialist and the switch from a four-man to a five-man rotation are the two biggest factors in this trend, which results in fewer complete games and fewer innings pitched among starters. It used to be that a superstar starter completed more than half of his starts; today, even pitchers such as Pedro Martinez rarely do. And when a pitcher of his magnitude does, it's generally when he's pitched a relatively economical game.

This piece by Don Malcolm at Baseball Primer suggests that managers and other baseball folks have been taking the research that's been done in the field over the past decade or so to heart. Malcolm compares a handful of recent seasons with regards to the distribution of games of 100+, 110+, 120+ and 130+ pitches. The results indicate a steady downturn in the number of high pitch count (120+) games. In 1988, according to Malcolm's numbers, around 20% of all games fell into this category. By 1998, it was below 14%, and last season around 11%.

Malcolm also compares this season's data with the first ten weeks of the other seasons he's examining. Early in the season, pitch counts tend to be lower; managers don't extend their pitchers to the max as often. This season, high pitch counts make up just 5.7% of all games, compared with around 10% during the same portion of last season, and over 15% in 1988. Malcolm suggests that the redefined strike zone may have something to do with this dramatic decrease, and illustrates a similar one year abberation from an earlier strike-zone change season—in this case, 1963.

The Brooklyn/Los Angeles Dodgers employed a statistician named Allan Roth from 1947 to 1964. Roth is considered to be one of the titans of baseball statistics. You've heard of On Base Percentage? That's his. Roth kept track of batter vs. pitcher breakdowns and is credited with helping Sandy Koufax to change his style of pitching based on his data. Way ahead of his time, Roth kept meticulous pitch logs of Dodgers games. More on him another time. Anyway, Malcolm summarizes Roth's data from the 1962-1964 seasons, and notes a dramatic dip in the 1963 season, when the strike zone was changed. High pitch count games went from 33% to about 26% and then back up to 34% during these three seasons (note how much higher these rates are than current ones). Malcolm proposes that this season may present a similar aberration.

Whether it is or not, there's no denying that pitch counts are definitely falling. The mere frequency of the phrase "pitch counts" on the lips of managers, players, agents, and even the mainstream baseball media shows that somebody has been paying attention. Somewhere amid the considerable amount of data is an idea whose time may have come.