Is Clay Buchholz’s change-up responsible for lefty struggles?

Boston Red Sox starting pitcher Clay Buchholz throws a pitch at Yankee Stadium in New York

Clay Buchholz is poised to take another step forward this year as a full-time member of the Red Sox rotation. He spent much of the first half of 2009 dominating Triple-A competition, and finally got his shot when Tim Wakefield experienced injury problems most of the second half.

His “emergence” wasn’t much of a surprise despite his godawful 6.75 ERA in 15 starts and one relief appearance during 2008, as his Fielding Independent Pitching was 4.69 and xFIP (FIP with home runs normalized to park and league data) 4.28 over 76 innings. Those respective FIPs were 4.69 and 4.09, covering 92 innings in 2009 with a 4.21 ERA all told.

An aspect to Buchholz’s game I don’t think has been made enough of, however, is his inability to pitch against left-handers. Okay, inability is a bit strong, so let’s go clear deficiency compared to right-handed batters.

Over 99 minor league innings for the PawSox in 2009, Buchholz posted a 2.36 ERA (3.37 FIP), 89 strikeouts and 30 walks — strong numbers, indeed.

Split it out to left- and right-handed data, and we see that Triple-A right-handers batted a scant .154 against Buchholz, with a 2.44 FIP. In the majors, that batting average went up to .228. There is no major-league FIP data against right-handers currently, but marked Buchholz’s “tOPS+” against right-handers at 90. tOPS+, as defined by B-R, is the OPS split relative to the overall OPS, with the lower number favoring the pitcher. Buchholz’s overall OPS+ was 100, so there’s a significant indication that Buchholz is quite adept against righties.

How about lefties? At the big-league level, batters raked to a .284 clip and tOPS+ of (naturally) 110. In the minor leagues, the FIP was 4.63 and batting average of .227.

Below is the left/right data of Buchholz over the last five years, split out between minor leagues and major leagues (major league data only available from 2007-2009). The minor league numbers have FIP, while the majors has tOPS+. I’ve left out batting average to reduce numbers overload, as they’re backed up by FIP and tOPS+ anyways.

Year Minors FIP L Minors FIP R Majors tOPS+ L Majors tOPS+ R
2005 2.20 3.29
2006 3.68 2.89
2007 2.89 2.12 139 40
2008 4.05 2.21 105 95
2009 4.63 2.44 110 90

The data just got a lot more interesting. While we can see that Buchholz has clearly struggled against lefties for two years now, he used to exhibit a much smaller split. Not only that, but it seems as if his splits showed up once he hit the majors in 2007, as he shows a marked right/left disparity despite no major disparity in the minor leagues.

The timing here of the sudden change in ineffectiveness against lefties in 2008 could possibly be traced to an arm slot issue. Prior to 2008, the Red Sox felt that his arm slot was a bit too high, which left fastballs (and presumably other pitches) hit harder. There may be something to Buchholz’s ineffectiveness against lefties stemming from an arm slot further away from his ear (the “lower” or further away from an ear than an arm slot, the better for the opposite-handed batter). There’s simply not enough data to say one way or another, so let’s look at something else.

Let’s turn to his pitch offerings for a moment to try to decipher what the issue may be.

Season Team FB SL CB CH XX PO
2007 Red Sox 52.9% (91.1) 6.5% (82.4) 15.4% (75.8) 25.2% (78.6) 0.30%
2008 Red Sox 47.4% (92.6) 7.6% (81.4) 21.4% (76.9) 23.6% (77.9) 0.40% 0.10%
2009 Red Sox 54.8% (93.5) 10.5% (85.8) 11.4% (77.7) 23.3% (80.1) 5.10%
Total * – – – 51.4% (92.9) 8.8% (83.9) 16.1% (77.1) 23.6% (79.0) 2.60% 0.10%

Your key: FB is fastball, SL is slider, CB is curveball, CH is changeup, XX is unknown pitch, PO is a pitchout. Nothing here stands out except major reliance on the change as his secondary pitch, and increased reliability on the curve. In terms of addressing performance against lefties, change-ups is the notable statistic here, which I’m about to explain.

In the 2010 Hardball Times Annual, there is an article by Dave Allen evaluating where pitches are predominantly thrown (“Where Was That Pitch?”), using pitch F/X. I’m about to delve into Buchholz’s pitch F/X and pitch data to try to find some sort of answer here using guidelines from Allen’s article plus Buchholz’s run values as shown here. (Briefly, to explain run values, think of a 95-mph fastball that was called for a strike. There is a small change — positive in the pitcher’s direction — of run expectancy. A positive change means the pitch was more effective at getting outs/strikes, while negative proves the batter took the pitch for a ride or a ball.)

(FT is a two-seamer, but there’s only four of that data over the entire 2009 season and Buchholz rarely uses it anymore. He has been known to throw a straight change as well as a circle change, which is tough to differentiate in pitch F/X.)

Firstly, using Allen’s information, we learn that across baseball change-ups are thrown most often in opposite-handed at-bats and have no platoon split, so we should consider most of Buchholz’s change-ups having been thrown to lefties. This is good, because we want to isolate why he’s bad against lefties and not righties. Also platoon-dependent are sliders, thrown mostly in same-handed at-bats and showing an extreme platoon split, meaning lefties would hammer these pitches if Buchholz threw them. Thus, sliders are avoided being thrown to left-handed batters.

Given that Buchholz’s slider run value in 2009 was 3.3 (and has had a positive value all three years) we can assume, then, that he does not throw sliders to left-handers. Let’s rule out sliders in trying to evaluate why Buchholz is lousy against lefties because they’re predominantly thrown to righties and sliders isn’t worthy of study in this article. Moving to change-ups, he had a -2.4 value this year but in years past, showed positive value (3.6, 2.0 in 2007 and 2008, respectively). Let’s table this for a moment.

Curveballs and fastballs do not have platoon splits and are thrown to any handedness of the batter with frequency. Using run values, Buchholz has minimally positive curve values in 2007 and 2009, and was a shade negative in 2008. The fastball was the major difference between his good years of 2007 and 2009 and bad year of 2008. His run value off fastballs in 2008 was an astounding -18.5, while the fastball was a negligible positive number in 2007 and 2009.

Now that we’ve gone through run values, let’s look at Buchholz’s pitch F/X locations of these pitches as modeled on the strike zone.

This is where Allen’s article comes in, where he mapped a pitch’s run value based on platoon splits (righty pitcher against righty hitter, etc.) and which location of a pitch was most effective. I’m about to compare Buchholz’s data to Allen’s. One caveat here is that Allen accounted for lefty/righty disparates, fusing the two sides of a plate into one. There is no such fusing in the image above. For example, a pitch inside to a left-handed batter as well as a right-handed one in Allen’s data would be in the same “location” allowing better data evaluation. In the image above, no such distinction has been made: an inside pitch to a lefty is on the left side of the image, while it’s on the right for a right-hander.

Let’s look at platoon-dependent data. Earlier in this article, it was said we should consider most of Buchholz’s change-ups as having been thrown to left-handers. We see that Buchholz predominantly threw the change-up from the middle of the plate to the inside (for a lefty) of the plate, sticking to the bottom of the strike zone, although there is an appreciable amount near the top. There are a smattering of pitches on the right side of the plate (which would be away from a lefty) but nothing major. He also pitches low and in to a right-handed batter. Again, most change-ups are thrown to opposite-handed batters so let’s assume those were thrown to lefties.

I’m going to use Allen’s run value data below, which has been scanned from the book (really, buy the book — it’s fantastic and is not a book geared to just statisticians, so don’t let that turn you off), to see if any of Buchholz’s change-ups are a red flag that lefty batters are jumping on.

The darker the circle, the lower the run value of the pitch. Here we see that change-ups that are outside and hitting the edge of the strike zone are the best change-ups a pitcher can throw. There doesn’t seem to be much difference between whether it’s inside and up or low, although there is a slight trend to it being low. Up and away, as well as low and away change-ups are good for the batter.  (Remember, Allen calibrated all data from right- and left-handed pitchers to be the same, so to look at this from a left-handed batter’s perspective, flip the graph over the y-axis.)

As Hans Landa in Inglorious Basterds would say, “That’s a BINGO!” While we have no hard data to look it up, what we’ve learned here so far about the change-up as it relates to Buchholz and left-handers is:

  • Buchholz has struggled against left-handed batters recently.
  • Of his repertoire, one specific pitch is thrown to opposite-handed batters with the most frequency.
  • Change-ups are predominantly thrown to opposite-handed batters.
  • Positive (good results for the pitcher) run values of change-ups trend towards those that are pitched away to a left-hander.
  • Using the 2009 pitch F/X data, we see that most of Buchholz’s change-ups were inside to a left-hander. A fair amount were also over the heart of the plate (bad) or low and away (also bad).

Based on this visual data, it’s clear to me that Buchholz’s choices of location for the change-up is why he suffered a negative run value of the change in 2009, and it wouldn’t surprise me in the slightest that this pitch is why he’s being tormented by lefties. How about his positive change-up values from 2007-2008? Let’s look at the two years of pitch data:

Just eyeballing this, it looks as if Buchholz stayed away from low and inside change-ups for the most part, which plays to left-handed batter’s strengths. He also threw more pitches that were away from the left-handed batter, which we’ve established from Allen’s graph helps a right-handed pitcher get left-handed batters out with change-ups. This seems to jive with previously-established data. (Interestingly enough, Buchholz said in a July 2007 interview that he thought his change-up was more effective against lefties… it seems as if that’s changed.)

Now, I want to caution that while we waded through a good amount of data and have learned more, there is more that can be done to evaluate Buchholz’s struggles against lefties. Some of my conclusions necessitated leaps of faith. The reasoning behind these leaps of faith is based in data, but they are leaps of faith regardless. Another caveat is that Buchholz uses the change so much in general, he may be the outlier in the data showing change-ups are predominantly thrown to lefties — but it wouldn’t shock me if lefties feasted on the change relative to Buchholz’s success against righties.

Are we closer to understanding why Buchholz struggles against lefties? I say we are. I say that pitching coach John Farrell and Buchholz should have a nice talk during spring training on where to throw change-ups to left-handed batters.