On Tuesday, July 2, one of the best hitters in baseball faced off against one of the best pitchers in baseball. Carlos Gonzalez, Colorado’s left fielder, had four at-bats against the Dodgers’ ace, Clayton Kershaw. Gonzalez went 0-for-3 with three strikeouts.
"Gonzalez, a left-handed hitter with a pretty defined platoon split, is still an excellent hitter against lefties when he swings at curveballs. So the play here is to keep it pretty fastball- and slider-heavy, as you’ll see from his slugging percentages against the three pitches Kershaw uses on lefties.
Gonzalez vs. LHP
Fourseam: .457 SLG
Slider: .436 SLG
Curveball: .573 SLG
So the play here is not to use the curveball much, maybe waste one or use it to keep the batter honest. Or, at least, it would be for most pitchers who have caught on. Kershaw has taken it to a different level. He has never, in their 109-pitch history against each other, thrown Gonzalez a curveball."
That, by the way, is no longer true. The number is all the way up to one after their most recent encounter.
The principle remains, however; the philosophy makes perfect sense. If Gonzalez performs the best against curveballs, throwing him fewer is the right call. After all, baseball is a simple game with a simple objective: get the batter out. And the best way to do it is stay away from his strengths.
More research has unveiled an interesting trend: the other Dodger left-handed pitchers show a similar pattern to Kershaw’s. The only other ones with more than ten meetings with Gonzalez are Ted Lilly and Chris Capuano. Capuano doesn’t throw curveballs much at all, so his information is irrelevant.
Lilly, though, has the same approach as Kershaw. In his career, about 9 percent of the pitches he throws to lefties have been curveballs. Against Gonzalez, that number drops all the way down to 3 percent.
Other Dodger lefties over the years have included Scott Elbert, Randy Wolf, Hong-Chih Kuo and now Hyun-Jin Ryu.
Elbert doesn’t have a curveball. In the three times Wolf faced Gonzalez while Wolf was a member of the Dodgers, he did not throw him a curveball. Kuo featured a curveball just about 4 percent of the time against lefties, but he too never threw one to Gonzalez. Ryu, despite throwing curves to lefties nearly 9 percent of the time, has also never thrown one to Gonzalez.
In general, data from pitcher-versus-hitter matchups is largely worthless because the sample sizes are too small to draw any viable conclusions. However, when we ignore the results—which is what we are doing—and focus on pitch selection, three plate appearances suddenly becomes 15 pitches or fifteen plate appearances becomes 50 pitches and we can start to get a sense of what the Dodgers are thinking: Gonzalez is a much better hitter against left-handed curveballs than he is against other pitches, so the Dodgers throw him fewer curves.
But Gonzalez does see curveballs about 13 percent of the time, so someone has to be throwing them. Who might that be?
Shockingly, it’s the San Francisco Giants, another team that faces Gonzalez several times per year. So, maybe, baseball isn’t so simple after all. Maybe baseball is weird, complicated, and convoluted, with multiple formulas for success. After all, both teams have extensive experience with Gonzalez, and yet each attacks him differently.
For the last few years, the Giants have featured a rotation with two lefties: Barry Zito, who throws his curveball nearly 20 percent of the time when facing left-handed batters, and Madison Bumgarner, who features his a little over 12 percent of the time.
Both Zito and Bumgarner have accumulated over 30 mound appearances against Gonzalez; in fact, they’re the lefties that Gonzalez has seen most often. And both pitchers throw curveballs to Gonzalez. 31 percent of the pitches Zito has thrown have been curves; 19 percent for Bumgarner.
Former Giants starter Jonathan Sanchez is also left-handed, and he faced Gonzalez 18 times. In those 18 matchups, he threw 65 pitches. 18 of those were curveballs, which comes out to 28 percent and is higher than his career mark of 21 percent.
Lefty specialist Jeremy Affeldt also fits this pattern. In 15 career appearances versus Gonzalez, 54 percent of his pitches were curveballs. This is substantially higher than the 33 percent that Affeldt has thrown throughout the rest of his career.
The Giants do have one final notable exception: left-handed specialist Javier Lopez. Lopez’s data does not line up exactly with the Giants’ other lefties (13 percent while with the Giants and 12 percent curveball usage against Gonzalez specifically), but his career mark is slightly misleading. Nearly half of the matchups between the two came in 2012, when Lopez used his curveball much less often that he has at any other point in his career.
Those four Giants pitchers (Bumgarner, Zito, Sanchez, and Affeldt) demonstrate a pattern directly in opposition to that of the Dodgers’ trio. The Giants have attacked Gonzalez with more curveballs than normal, while the Dodgers, acknowledging Gonzalez’s prowess against curves, throw him fewer.
Of course, it’s impossible to fully understand this information from the small sample sizes available. Batter/pitcher matchups are notoriously unreliable, and—as Yasiel Puig’s .443/.473/.745 line in 112 plate appearances has shown—a hitter’s results in a small number of plate appearances don't necessarily tell us much about his effectiveness in the given situation.
However, this information is more useful when analyzing the process behind pitch selection and sequencing, which is something we know comparatively little about. Catcher defense has long been difficult for analysts to quantify, primarily because so much of what they do is based on counterfactuals.
A debate also exists about who gets what credit for pitcher performance. Catcher ERA has long been discredited, but “the mystique of being a defensive stalwart” lives on. We don’t really know how or why pitchers throw the pitches they do, at least beyond the basics of the catcher putting the signs down and the pitcher gets to either nod his approval or shake for a different option.
In a recent BP chat, the aforementioned Zachary Levine—formerly a beat reporter covering the Houston Astros—mentioned “a meeting every day about two hours before the game featuring the starting pitcher, catcher and pitching coach where they go over every hitter.” That this occurs is no secret; teams have to discuss their plans for attacking each member of the opposing lineup.
However, logic dictates that many of them would come to the same conclusion—i.e., Carlos Gonzalez hits curveballs well so we’ll throw him fewer. But, at least in the case of the Dodgers and Giants, this is not the case.
One obvious explanation does not hold up to any level of scrutiny: perhaps the Giants are encouraging their pitchers to pitch to their strengths, while the Dodgers are advocating pitching to a hitter’s weakness. Bumgarner allows lefties to hit .247 against his curveball, though, which is the highest of any of his pitches.
The next obvious question that comes to mind then is where this discrepancy comes from and why it exists. Clearly, each team has a different process by which it arrives at its daily “plan.” But the extent to which these processes differ is startling.
The role of pitching coaches is debated, and Russell Carleton explored this topic at Baseball Prospectus earlier this year by calculating the effects of specific pitching coaches on a staff’s performance. His article explores the “what” of pitching coaches’ results, but we don’t have much information on the “why.”
A possible conclusion from the above data centers on the pitching coach’s philosophy and methodology of attacking hitters. However, there’s still too much unknown about what happens in those pregame meetings and just how much authority individual pitchers and catchers have in determining a strategy for us to get a full picture.
But as a potential vehicle for evaluating the process of pitching rather than simply the results, pitch sequencing and usage remain valuable commodities.