Fun with pitch distributions: Gallardo, Lohse, and Estrada

EDITOR'S NOTE: The Frosty Mug is off today. It will return tomorrow. - KL

Last week I wrote up a fairly lengthy scouting report on Travis Wood of the Cubs for a series preview, and when I finished I heard Joe Block through the radio inform me that the Cubs were going to bump back Wood, a lefty, so he would not have to face the right-handed heavy Brewers. My report was heavy on splits, and how Wood completely altered his pitch selection against righties and lefties, as many pitchers do. It occurred to me at that time that I knew more about the handedness-dependent pitch selection of Travis Wood than of any of the Brewer pitchers. So join me as I learn a bit about the pitch selections of the players we probably would consider the top 3 in the rotation (with graphs!).

Yovani Gallardo

Over time Yovani's overall profile has changed. When he came into the league the curveball was his primary offspeed pitch, but the slider has overtaken the curve, and the changeup has also fallen from being about 5-6% of his pitches to only 3 total changeups so far in his first four starts 2013. Here's his distribution graph from last season, showing percentages of times he threw each pitch in the following situations: against lefties overall (including with 2 strikes), against lefties only in 2 strike counts, against righties overall (including with 2 strikes), against righties with 2 strike counts, and just overall pitch distribution including everything.

**All of these graphs are data from 2012**, which I think is the most fair way to approximate current trends.

If there's a very generalized rule to pitching, it's changeups/splitters to batters on the opposite side of the plate, sliders/curves to those on your side of the plate. Gallardo, as a guy who does not rely on the changeup much at all anymore, is a righthander whose preferences across batters are relatively consistent. Like nearly all pitchers he relies more on off-speed pitches with 2 strikes to generate Ks, but he does so than a lesser extent than most-- and it's possible that some of his ability to keep batters off balance is due to his relative use of the fastball in 2-strike counts.

On the changeup-- it's simple to say he should throw it more, but it's a more complicated situation than that. According to Fangraphs it is, on average, the least effective pitch on a rate basis that he has thrown over the course of his career. Gallardo was worse than the league average split in that situation in the past 2 seasons. It's very possible that throwing it more often could make it a more effective pitch, improve his performance against lefties, and help him to improve overall, and that may be something he should look into if he wants to make the step to the next level by improving his performance against lefties.

Kyle Lohse

Lohse isn't a guy that has changed a lot about his repertoire at this stage in his career, but he has relied more on his changeup in recent years. Most encouragingly, he has increased his swinging strike percentage on the changeup every year since we have had data in 2007, from about 13% to about 17% now. He also uses his breaking stuff to get ground balls-- his slider, curve, and change all average higher rates of ground balls than his regular sinker, though last year the changeup was his most likely pitch to be a hit for a home run (albeit by a very small amount).

As I mentioned, Lohse is a much more traditional pitcher in the sense that he uses his changeup as the primary weapon against lefties and the slider as the primary weapon against righties. That 2-strike row for lefties shows that nearly 40% of the strikeout pitches are changeups, quite a high number compared to other pitchers. Lohse is more of a balanced pitcher who, though he is better against righties and does not have a reverse split, beat the league average OPS against by about the same amount in the RHP vs RHB and RHP vs LHB categories last season. That's been the trend throughout his career, though he does not always beat the league averages overall as much as he did last year when he went 16-3*, he tends to be pretty balanced in comparison to league averages on each side of the plate.

Marco Estrada

Estrada is primarily a 3-pitch guy, he mixes in an occasional cutter with his primary fastballs but the amount is <1% so I mixed them in with the others. As a curve/change guy he's an even more straighforward case study in the idea that you only need a fastball and the one offspeed pitch for each side of the plate. It's possible that Marco's effectiveness has been due in part to the fact that he does actually mix the unanticipated pitches to the opposite side on a regular basis.

As you can see, he goes more to the curveball against righties and the changeup against lefties but that graph doesn't look much Lohse's. Estrada, too, was about equally effective in relative terms against righties and lefties, beating the league averages by about the same amount on each side (though he was more effective against righties than lefties overall). We have far less data on Estrada than anybody else but one thing to watch this year will be too see if that trend continues, or if he will have to alter his game plan with time to stay ahead of lefthanders. Like Lohse, Estrada's best swing-and-miss pitch is his changeup, but in contrast it's more likely to be hit in the air.

When talking about these relative splits it's important to understand that the distinction has nothing to do with how good each pitcher is overall, but how on balance their production varies against righties and lefties. After all, right-handed pitcher who gives up a .600 OPS to lefties and a .700 OPS to righties is better than one who gives up an .800 OPS to each side. What we have here are 3 very different process/results combinations-- Gallardo has balanced selection but gets hit relatively harder by lefties than righties, Lohse varies his approaches to each side and is about even in comparison to league averages on each side, and Estrada is more like Gallardo with the balance in pitch selection but is also relatively equal on each side of the plate. These factors don't tell us how good a pitcher is overall, but might give us an idea of where there is room to improve.

Nothing in this post should probably shock an avid Brewers viewer, but it's interesting to see it in graphical form. The far more interesting aspect of pitch selection is the game theory behind throwing each individual pitch, but sadly that data is difficult to aggregate and present in a way that we can see long-term trends. So enjoy the graphs, and at least now I know as much about the Brewers staff as I do about stupid Travis Wood.

*He almost never does this.

Data from Brooks Baseball, Fangraphs, and Baseball-Reference.