Carpenter’s Pitch Selection

Sports Illustrated’s Tom Verducci chronicled Chris Carpenter’s transformation this season in an article published during St. Louis’ World Series run. In it, the Cardinals’ ace credited pitching coach Dave Duncan with his professional turnaround. The premise of Verducci’s story was a simple one: Carpenter — then with Toronto — began his career as a hard-thrower with a four-seam fastball, curveball and not much in the way of command. But after arriving in St. Louis prior to the 2004 season — voilà! — Carpenter learned a two-seam fastball and Duncan showed him how to command the bottom of the strike zone. In Carpenter’s words:

The classic Dunc game: keep the ball at the bottom of the strike zone and you’re going to make it harder for the other team to score. It’s not so much about trying to hit corners as it is pounding the bottom of the zone with movement.

With this in mind, it’s no surprise that, for 2009 and 2010, Carpenter threw almost all of his fastballs with a sinker grip. He also threw a cutter, a curveball and a change — but we won’t deal with the curve and the change much in this post. The images below show spin-deflection charts for his fastballs and cutters from some of his 2009 and 2010 starts so that we can see his pitch movement, on a game level.

(click to enlarge)

You can see that, in these examples, there are a lot of pitches that move laterally about 8 to 10 inches – those are the fastballs with the two-seam grip that Duncan taught Carpenter. Some starts have pitches that look a little bit straighter, with less sink and tail to them. These are the four-seam fastballs, which look rare, based on the games selected above. The pitches on the other side of the axis are cut fastballs, which move away from a right-handed hitter.

Now let’s look at those same charts for three 2011 starts.

The four-seam cluster is now very prominent. Tracking the progression of his pitch selection confirms that Carpenter’s reliance on the four-seam fastball was new in 2011, and it really started to become prominent by the end of May.

The next thing I wanted to examine was the effect that the ball/strike count had on Carpenter’s pitch selection. My apologies if the graph below looks a little bit funky: it compares Carpenter’s 2011 pitch selection to his 2010 pitch selection, split up by each of the 12 ball/strike permutations (which are sorted by increasing run expectancy). Positive numbers show an increase in usage from 2010; negative numbers show a decrease. Particularly against lefties, Carpenter’s pitch selection has changed the most on two-strike and in 3-0 counts.

So Carpenter turned himself into a different pitcher in 2011. But what kinds of implications did his change in pitch mix have? The following table has some generic PITCHf/x metrics on the two-seam and four-seam fastballs over the 2009 through 2011 seasons.

            #        mph      Swing    Whiff    Ball     Zone     GB%
Two-Seam    4219     92.0     .448     .093     .329     .523     .544
Four-Seam   755      92.7     .465     .194     .331     .499     .333

Two-seam fastballs typically get grounders but not many whiffs; four-seamers typically get whiffs but not many grounders. Carpenter’s pitches are no exception. In fact, his two-seamer is better than average (~53%) at getting grounders, and his four-seamer is better than average (~17%) at getting whiffs. Basically, you could say that they’re both “plus” pitches for their respective purposes and relative to pitch type.

Overall, Carpenter was much less of a ground-ball pitcher in 2011 than he has been throughout his career in St. Louis. Dating back to 2004, he’s been at ~52% ground-balls-per-pitch-in-play, but that rate dropped to only 45% this year (including postseason starts). His strikeout rate — including postseason starts — was just under 7K/9, which is where he’s been in the past three years. The biggest difference with Carpenter’s performance this year was his lack of grounders — and this can be explained by the sudden increase in four-seamers. It’s true that his new pitch was better at getting swinging strikes — compared to his sinker — but his curve and cutter are better than both. So if the four-seam fastball is taking the place of curves and cutters in two-strike counts, he’s putting himself at a disadvantage, particularly because the pitch yields an extreme number of fly balls. The evolution of Carpenter’s repertoire is puzzling, especially considering how much credit he’s given to Dave Duncan for kick-starting his career.

PITCHf/x data are from MLB Advanced Media and here are courtesy of Joe Lefkowitz’s tool.  (Two of Carpenter’s games since 2009 are missing in my dataset.) 

A note on the strike zone: the definition of the strike zone here is the one Mike Fast adopted in his Baseball Prospectus articles, “The Real Strike Zone, Part 2” and “A Zone of Their Own.”  The vertical zone is fixed from 1.75 feet above the ground to 3.4 feet above the ground. The horizontal zone to LHB is 1.2 feet inside from the center of the plate to 0.81 feet outside; and the horizontal zone to RHB is 1.03 feet inside from the center of the plate to 1 foot outside.