Clay Buchholz is a polarizing pitcher. Sometimes he’s really good and sometimes he’s really bad. When faced with adversity he can show a lack of awareness. Ian York takes a look at what happens to Buchholz with men on base.
Clay Buchholz has had maddeningly inconsistent results throughout his career. Some games he shows ace potential; in others, he gets chased in under three innings after giving up five runs.
In 2015, Buchholz has a profoundly unimpressive 5.73 ERA, placing him 103rd of 110 qualified pitchers. On the other hand, looking at his peripherals shows some reason for optimism. His FIP is a much more respectable 3.10, slotting him in 22nd place. And his BABIP (batting average on balls in play) of .393 is the second-worst among qualified starters. BABIP is a statistic that can be looked at differently for pitchers and hitters. Hitters, to a certain extent, have some control over their BABIP. A player who regularly hits hard line drives will have a higher BABIP than one who hits soft ground balls, and this difference tends to be consistent and reproducible from one year to the next. BABIP can still give an estimate of a batter’s luck or lack thereof, because a very high BABIP (say, over .400) is certainly an indicator that the batter got some lucky hits; but a player who has a moderately high BABIP (say .350) over a long enough period is more than just lucky.
That seems to be less true for pitchers. Pitchers have little control over what happens to a hit pitch, so a pitcher with even a moderately high BABIP is very likely to have been purely unlucky, while one with a very low BABIP probably didn’t get that by his pitching alone, but through the help of good luck and good fielding.
Still, it may be interesting to look more deeply into Buchholz’s BABIP. “Sampo Gida”, on the sonsofsamhorn.net forum, pointed out that Buchholz was being “hammered” with runners in scoring position. This struck a chord, because one very subjective observation about Buchholz is that he really doesn’t seem to like having runners on base. As soon as a man reaches first, Buchholz slows down, throws to first repeatedly, and generally seems like a different pitcher.
Sure enough, with men on, Buchholz has had an abysmal BABIP of .462, while with the bases empty his BABIP was just .333. This was calculated by looking at Buchholz’s BABIP through his first seven games of 2015 with PITCHf/x data with and without runners on base. By comparison, the average BABIP for all pitchers with runners on base is .297, and with bases empty it is .290.
Conventional wisdom says that pitchers have no control over their BABIP, so that a high BABIP is not predictive. In other words, given a reasonable time period (certainly more than seven games) a very high BABIP is almost sure to revert back toward the league average. However, it seems possible that a very poor pitcher who gives up many hard-hit pitches may “really” have a higher BABIP because he is giving up quality contact, rather than simply being unlucky. Is the 2015 version of Buchholz with men on base this hypothetical very poor pitcher? Is he truly being unlucky, or is it possible he has pitched poorly enough with men on base to “deserve” the higher BABIP?
If the latter was the case, then there should be some other factors we can point to that support the “feet of Clay” theory. For example, if he has a history of a large situational BABIP discrepancy , then this year’s small sample size would become more significant. I looked at the five years previous to this one:
|Year||BABIP||BABIP (bases empty)||BABIP (runners on)|
(These numbers include all the games for which PITCHf/x has information, so they differ slightly from baseball-reference.com’s regular-season game numbers.)
If anything, historically Buchholz has had slightly lower BABIP with men on base. On the other hand, last year he did have a .100-point higher BABIP with men on, so the possibility remains open that this is an actual trend.
Is there any evidence that Buchholz pitches differently in different situations? For example, with men on base, might Buchholz throw different (more hittable) pitches, or may he throw a slower fastball or a curve with less break to it, or may he lose location and throw pitches down the heart of the plate?
Looking at the mix of pitch types, it is clear that he (or his catchers) use a slightly different set of pitches with men on base:
With men on base, Buchholz throws fewer curves, and makes up the difference with cut fastballs. This seems like a reasonable approach; the cutter breaks in the same direction as the curve, but is significantly faster, so it offers less time for runners to steal. He has shown this same tendency for several years. Is it possible that Buchholz is curbing steals, but offering a more hittable pitch with the cutter, resulting in a higher BABIP?
It is possible to answer this question by looking at the BABIP for each pitch in different situations. At this point, it is very important to emphasize that we are now dealing with very small numbers. For 2015, we are chopping seven games into two different game situations (men on, bases empty) and then further subdividing these into the results for five different pitch types. With these sample sizes, one seeing-eye bloop hit can easily double BABIP. It’s unlikely that any of these points are anything other than chance, but we can still look at the numbers and see if anything might be worth following up on as the data get larger (which is the cheap pine veneer of rationalization covering the particleboard of small sample size).
Here is a summary of the number of hits per pitch and the BABIP for Buchholz’s pitches in 2015 in different situations. The biggest difference in BABIP (as well as hittability) comes from the curve, with the changeup and fastball also showing some increase with men on base:
Historically, Buchholz’s BABIP has been different for some pitches with bases empty vs. with men on, but there is no obvious pattern:
|Pitch type||Empty||Men on||Empty||Men on||Empty||Men on|
In 2015, the cutter has a better BABIP with men on base than the curve it replaces. (Before becoming too enamored with the 0.000 BABIP for some pitches, remember: BABIP does not count home runs.) Again, this striking difference in hits per pitch and BABIP is not seen across baseball in general:
The mix of pitches with men on doesn’t seem to easily explain the increased BABIP. We can ask more about the quality of the pitches by looking at the three major characteristics of a pitch — speed, break angle, and break length. Each pitch cluster for the bases-empty situation is approximately outlined, and the same outlines are shown on the men-on-base situation:
There is little or no difference in pitch velocity or break with men on base or bases empty, with the only possible exception being that his four-seam fastball may break at a slightly different angle with men on base.
Finally, we can look at the pitch location. Whether intentionally or not, does Buchholz position his pitches differently when men are on base? Here we can compare each pitch: to right- or left-handed batters, with men on or with the bases empty. (Again, it is important to remember that this chops up sample size still further by adding batter handedness to the combinations.) In each of these charts, the pitches thrown to a left-handed batter are shown on top, and to RHB on the bottom; pitches with bases empty are on the left, with men on base to the right. The 2014 de facto strike zone is shown as the dashed oblong shape. The pitches with the biggest increase in BABIP are Buchholz’s changeup and his curve. Given the small number of pitches, these locations are very similar:
Buchholz’s changeup shows quite a difference in location to left-handed batters, but that is not linked to differences in BABIP. None of the LHB hit the pitch in either location. In fact, the apparent effect of increased BABIP here is due to the small sample size. It is linked to one extra hit by a RHB with men on base:
We look at peripherals like FIP and BABIP to judge whether a pitcher has been unlucky or whether poor outcomes reflect poor pitching, which may help predict his results going forward. Buchholz’s FIP, and its very large difference from his ERA, suggests that he may have been pitching reasonably well but having bad luck (or bad fielding) that led to bad outcomes. His high BABIP suggests the same thing. On the other hand, it seemed at least theoretically possible there might be some genuine underlying factor driving his very high BABIP with runners on base. Given that his pitches seem very similar whether men are on or not, we are left with three possible explanations:
- Buchholz might be doing something different with men on base. For example, he could be tipping his pitches. This is possible, but not likely.
- We could have simply reversed the direction of causality. Rather than having a high BABIP because runners are on base, it could be that runners get on base when Buchholz has a high BABIP. This is difficult to test, and may well be a part of the story, but the way we are combining pitches from different games makes that a little less likely.
- Buchholz has been pitching reasonably well, but he has simply been unlucky with men on base. With the small sample size, one or two extra hits falling in have a massive effect on his overall numbers. This is by far the most likely explanation.
As frustrating as it has been to see Buchholz get hammered this year, the most likely explanation is not that he has become a terrible pitcher, but that he has been exceptionally unlucky so far, and/or that the defense behind him has been terrible. To the extent that he has been unlucky, we can reasonably expect him to improve quite a bit over the rest of the season.