X Pitch Pitchers

In last year’s spring training, Trevor Bauer made headlines by telling the batter what was coming. As is typical in bullpen sessions, he communicated through glove motions in what is essentially a universal language in baseball. He did manage to record an out, but needless to say such an experiment would be silly in the regular season… let’s talk about Jake McGee.

McGee has always been fastball heavy. In 2014, he threw 96.3% fastballs, but they were a balanced mix of four-seam and two-seam fastballs. He did not mix up speeds much, but was able to keep batters off balance changing location and movement. From 2015 to 2019, he ditched the two-seamer for a slider with mixed results. The slider did alright, but his overall success as a pitcher was falling apart. In 2020, after being released by the Rockies, the Dodgers gave him a chance with an unprecedented game plan: Just throw four-seam fastballs.

I am not going to throw a percentage at you because I think it understates the point here. McGee threw 332 pitches in 2020. Ten were sliders. TEN. It is not like he was nibbling the zone with elite precision, he was just pounding the zone and betting no one could do damage. It worked. He posted a 2.66 ERA with most ERA indicators suggesting he was unlucky (somehow).

What an amazing case study of a true ‘one pitch pitcher’. He will be someone to watch next year, but there is no way he could succeed as a starting pitcher. McGee actually came up in the minor leagues as a starter, but was moved to the bullpen due to his lack of secondary pitches. To survive multiple times through the batting order, you need a deeper bag of tricks to keep hitters guessing. Right? I thought it would be fun to look at the biggest ‘X pitch pitchers’ of the statcast era. Starters may not be as extreme as McGee, but I am sure each case study will have their own interesting story to tell. So without further ado, which starting pitcher is the real biggest one pitch pitcher?

I miss knuckleballs. It feels like we see more from position players than actual pitchers nowadays. It makes sense given the current state of the MLB: Why nurture a volatile knuckleballer when you have a farm system full of kids pumping gas?

Regardless, Dickey makes sense as the truest one pitch pitcher in recent memory among starters. He had no idea where it was going and neither did hitters. Therefore, it did not matter if the opposition knew what was coming. They did not know how it was coming.

Dickey’s supremacy as the top one pitch starter is unquestioned – he was second and third on the list too – but he does benefit from how we classify pitches. As I have discussed before, one pitch type throw at two speeds is effectively two pitches. While Dickey mixed in other offerings, mixing up knuckleball speeds furthered his ability to make hitters uncomfortable and survive as a one pitch pitcher.

Remember when Chris Archer was an all-star? Those were simpler times. A year before the Rays introduced the baseball world to the Opener-Headliner pairing, Archer presented himself as a perfect Headliner option. He had a spectacular fastball-slider pairing, but not much else. As traditional baseball thinking would suggest, he accordingly struggled to go deep into games. He was dominant two times through the order, but once he got around 75 pitches, he would unravel.

The opener strategy is built for teams with deep bullpens and aiding pitchers who struggle with that third time through. Perhaps it could have accentuated his strengths bringing his career to new heights. However, as Pirates fans are well aware, that is not how things wound up. Pittsburgh acquired Archer in one of the most lopsided trades in MLB history and added a curveball and sinker to his arsenal.

These additions might have improved his third time through stats (0.698 OPS in 2019), but killed his effectiveness the first two times through the order (0.827 OPS in 2019). Archer might have been ailing, but the newly added sinker was struggling. Having just been re-acquired by the Rays, we will have to see what the Tampa Bay Pitcher-Whisperers decide to change.

In 2016, McHugh threw almost identical percentages of his top three pitches. Even on the first pitches, all three were within three usage percentage points of each other. No matter the count, it was quite the guessing game for hitters.

Although he has traded in his curveball for an effective slider, his search for other quality pitches haven’t been as successful. His sinker has had the highest xwOBA by far of any of his pitches since 2016, and in his last full season as a starter his changeup wasn’t much better. 

McHugh signed with the Rays this offseason, which means he is going to be a Cy Young candidate this year. In all seriousness, he has started less than 9% of his games since 2018. As a strong swingman candidate in 2021, the need for fourth and fifth pitches is reduced. Perhaps he will return to his three pronged approach.

In one of the most anti-sinker eras of baseball history, Stephen Strasburg nearly won a Cy Young by tripling his sinker usage. More accurately, he traded 4-seam and slider usage for curveballs and sinkers in 2019. Along with earning him the ‘four pitch pitcher’ crown, this adjustment gave his offspeed the arsenal environment to shine.

Improvements like these always fascinate me as they speak to the counter-intuitive nature of pitching’s game theory. In both 2017 and 2018, Strasburg’s sinker was his worst pitch in terms of run value. His slider may not have been elite, but its run value was certainly better than his sinker’s. Despite this, throwing more sinkers proved to be valuable because of what it did for his other pitches.

This case study reminds me of one of my favorite game theory examples: Corey Kluber’s sinker. According to Fangraphs’ pitch info values, it has never been a positive pitch for him. Throughout Kluber’s prolific career, his primary pitch has always translated to below average results. He was willing to offer it up to hitters because it let his cutter and curveball thrive.

The four pitch pitcher approach worked for Strasburg in 2019, but his one 2020 start hinted he was moving away from it. In 85 pitches he threw 38 4-seamers and one sinker. If he can stay healthy, Straburg might just be coming for Collin McHugh’s ‘three pitch pitcher’ crown in 2021.

Shields was always a very balanced pitcher. He threw each of his top five pitches at least 10% of the time every season. As crazy as that is, it wasn’t even too special for him. His 2016 season was also in contention for the ‘five pitch pitcher’ crown. 

Almost all of the guys at the top of the five pitch pitchers list seemed to fit the same archetype. First, most of them had below average fastball velocities. That played into the fact that they all had a 4-seamer and 2-seamer, and most of them threw a cutter. Splitting the total fastball percentage up into two or three, makes each specific variation closer to a normal secondary pitch.

 Shields had a good career, but a surprisingly tragic ending. His rapid decline started right after this 2015 season. His velocity decreased by around 1 mph each year, which led to his strikeout rate plummeting and hard hit rate skyrocketing. Combine that with the fact that the Royals won the World Series in their first year without him, he gave up a home run to Bartolo Colon, and was traded for Fernando Tatis, and you almost feel bad for the guy who made 114 million dollars.

Although Miley struggled mightily in the 2017 season, it helped the former all-star’s career get back on track. It was the first year that he started replacing some of his 4-seam fastballs and sinkers with cutters, and it worked. Due to lower exit velocities and launch angles, and a much higher whiff rate, his cutter had a far better xwOBA than either of his fastballs. 

After the 2017 season, Miley decided to make the cutter his primary pitch. He upped its usage from around 14% to over 40%, and in 2020 he was at 50%. The results followed as well. In 2016 – when he didn’t throw cutters at all – and in 2017 his xERA was 5.15 and 5.70 respectively. Once he made it his primary pitch he had two consecutive years with an xERA under four. 

In the shortened 2020 season, Miley pitched only fourteen innings, but may have given us a glimpse of what is coming next. As I already mentioned, he threw his cutter more than ever before, but it may have been a little different. He threw them almost two mph slower than his career average, but also had more movement on it than ever before. The sample size clearly wasn’t large enough to determine if it can be effective, but Miley clearly understands how important that pitch is for him.

Sliders were technically his seventh most used pitch, but I refuse to pass the opportunity to point out Odrisamer Despaigne threw 160 eephuses. Eephusi? ONE HUNDRED SIXTY EEPHUSI. He did a great job at just placing them in the zone, but probably should’ve used them more as a ‘get me over’ instead of using them so often on two-strike counts. They only had a 8.3 PutAway% in 2015 and he ditched the pitch – or at least its regular usage – the next year.

It is a shame we never got to see more of Despaigne, but that was due to natural selection. While he lasted until 2019, he never mustered an ERA below 4.00 after his rookie season. His velocity climbed and he had a deep bag of tricks, just not what it took to consistently succeed in the MLB. Despaigne currently resides on a KBO roster, so we will have to get our eephus fix elsewhere. For now, the occasional Greinke lob will have to suffice until Despaigne makes a comeback.

Darvish throws just about every pitch known to man. Nobody else really throws more than seven pitches, so the sheer size of his arsenal is what brings him to the top of the list. The thing is though, he’s even more balanced than Baseball Savant gives him credit for. As I pointed out in my tweet he threw some misidentified sinkers that were really changeups or splitters.

What’s even crazier about Darvish is that you could argue he has more than ten pitches. In a recent pitch velocity article I pointed out that some of his pitches have a bimodal distribution, meaning that he almost has two different types of the same pitch. The baseball community could easily give him credit for his hard cutter, slow slider, and any of his different curveballs, but then what would we name them?

Darvish doesn’t seem to have any loyalties to any pitch either. His pitch usage seems to criss cross every year in a similar way that the xwOBA for his individual pitches do. The level of variation and unpredictability that Darvish has is unmatched by anyone of his generation. So sit back, watch, and enjoy it, because you never know when we’ll have someone like this again.

For each category, the crown was given to whichever pitcher had the highest count of the X most used pitch in their arsenal. This gave a good mix of incevizing high usage rates and high total pitch counts – meaning our list would be full of starters, not dominated by random relievers.

With our lineup of top candidates, who takes the crown of having the most balanced arsenal? Is it one of the mid-range pitchers with nearly equal mixes of offerings (Shields/Miley) or someone with a deeper bag of tricks (Despaigne/Darvish)? To answer this wider-lens question, I used the Herfindahl-Hirschman Index (HHI). This market concentration metric takes the sum of the squared percentages to gauge how balanced the market – or in this case pitch repertoire – is. If monopolistic, the HHI will be near 10000. If there are a lot of firms (pitches) with similar market shares (usages), the metric will be closer to 0.

The lowest score from the statcast era was 1779 belonging to… Mat Latos? In 2016, his arsenal featured a 4-seamer (28.3%), slider (18.3%), splitter (16.0%), cutter (13.4%), sinker (12.1%), changeup (6.5%), and curveball (5.5%). The fact HHI squares the percentages makes it key his top few pitches are so low in usage. Being an ‘eight pitch pitcher’ was only about the eighth pitch; being the most balanced widens the lens to the whole arsenal.

The crown of most balanced pitcher did not do much for Mat Latos. After a dominant five year run to begin the decade, his strikeout and walk rates started turning the wrong directions. The velocity was gone and Latos was grasping at straws.

Having a balanced mix of pitches sounds good, especially for a starter, but often leans away from optimality. So to the Chris Archer’s and Yu Darvish’s of the world – embrace your weirdness. It is what makes you special and gives nerds like me something to rave about.