When a Limited Repertoire’s a Good-Enough Repertoire

I was reading something about Clayton Kershaw the other day. Kershaw’s been the guy to write and read about, because he’s been amazing, and because he just got paid, and because there sure as shoot hasn’t been anything else happening around the league. This is the article, and it talks about Kershaw’s ongoing search for a reliable changeup, to go along with all of his other weapons. A particular section:

And so I am boundlessly curious about the changeup, the mystical less-is-more pitch, and Kershaw’s curiosity for it, and the evolution of Kershaw as a pitcher and if there is room for another pitch along the way.

Pitchers like to say the hitters will decide. So far, hitters aren’t suggesting Kershaw needs another pitch. One or two fewer, maybe.

Read it over quickly and you might dismiss that as a worthless old bromide. But there’s a lot of truth tucked away in there — hitters will let pitchers know when they need to make changes. If a pitcher isn’t having problems, the pitcher’s having success, and success isn’t something you mess with unless it gets threatened. Kershaw’s facing no threat, but then this isn’t about him.

There are young starting pitchers in baseball right now. A whole mess of ‘em. One of them goes by the name of Michael Wacha. This is about him, and some others. Shelby Miller. Tony Cingrani. Sonny Gray. Chris Archer. A certain pitcher-type in general. From those five specific starters, the worst major-league ERA- last year was 85. The worst xFIP- last year was 100. It’s probably time to say something about limited repertoires.

People ask about these guys in chats all the time. Wacha and Cingrani especially, it seems like. And they ask for a particular reason. According to conventional wisdom, in order to be an effective starting pitcher in the major leagues, it’s all but mandatory to have three pitches. Three reliable pitches, that is, and if you’ve got a fourth one, all the better. Wacha’s got two reliable pitches. Miller’s got two reliable pitches. Cingrani might just have one, to be honest. Gray lives off his fastball and his curve. Archer lives off his fastball and his slider. So, fans are concerned, and curious — what does the future hold for these guys? How important would it be, really, for them to fill out their respective arsenals?

The conventional wisdom makes some sense, absolutely. Relievers usually have two pitches. Having a third pitch allows the starting pitcher to give the hitters different looks, from both sides of the plate, in any inning. Over a long outing you have to avoid being predictable, and so forth. But we might as well do what we can with the numbers we have, to see how they shake out. And it just so happens we have pitch-type data ranging all the way back a dozen years.

So I collected that data for all starting pitchers over the past dozen years, season by season. I made note of their ages, and I eliminated knuckleballers, since they’re complete and utter freaks. At that point, I settled on a 10% frequency cutoff. That is, I considered a pitch type reliable if the pitcher in question threw it at least 10% of the time. It’s somewhat arbitrary, but it feels good enough to me to work. For each pitcher I calculated the number of reliable pitches, and then I grouped all the player-seasons by age. How do the one- and two-pitch starters come out looking?

  • Through 25: 26% of WAR in 27% of the innings
  • Ages 26-30: 26% of WAR in 25% of the innings
  • Ages 31-35: 23% of WAR in 22% of the innings
  • 36 and up: 44% of WAR in 37% of the innings

Things get kind of weird in the 30s, perhaps owing to smaller sample sizes. If you just combine the last two into a group of starters 31 and older, then one- and two-pitch starters account for 30% of WAR in 26% of the innings. Those limited starters who remain are still able to have success, and if it becomes more important to have a broader repertoire over time, as you lose velocity, that doesn’t show up here. It seems like the limited starters do just fine, relative to the rest of the starters.

There’s a considerable history of two-pitch starters making it and lasting a while, despite the on-paper disadvantage. Ervin Santana, right now, is an upper-tier free agent. Randy Johnson did almost everything with his fastball and slider, although I’ll grant that Johnson was a freak of his own. Tom Glavine loved his changeup and never really developed a reliable breaking ball. Ben Sheets had a fastball and a curve and basically nothing else. David Wells was similar. Jason Schmidt preferred to pair a changeup with his fastball. Brandon Webb leaned heavily on his sinker and curve. Bartolo Colon makes headlines when he doesn’t throw some sort of fastball. Rich Harden wound up a fastball/changeup guy. It goes on, and while these guys aren’t the norm, they’re common enough to have established a precedent. Baseball’s had a lot of successful two-pitch starters. It’s even had successful more or less one-pitch starters.

The gist being: you can make it with a limited repertoire. It just has to be a good limited repertoire. The fewer pitches a guy throws, the smaller might be his margin of error, but as noted before, hitters will send signals when the pitcher’s in trouble, and if a pitcher isn’t in trouble, he’s doing well enough. Which isn’t to say pitchers can’t try to improve, and it probably couldn’t hurt to have an expanded repertoire, but if opposing batters struggle to do damage, that’s the most meaningful thing. At the end of the day, pitchers are just trying to get outs, and they don’t get bonus points for getting outs with the broadest pitch variety.

Everything being equal, it’s presumably better to have more pitches. But it’s important to keep in mind that when we’re looking at the major leagues like this, everything isn’t equal. The starters who make it up with fewer pitches are selected for being in possession of particularly good pitches, pitches that allowed them to have success at every lower level. They would’ve had to prove themselves as starters, and odds are, each pitch has a higher average quality rating than each pitch belonging to a guy with a greater number of pitches. Wacha, for example, was selected because the Cardinals believed he could succeed out of the rotation, despite really just having a fastball and change. And then Wacha, sure enough, had good initial success. If a guy’s in the majors and starting with two reliable pitches, that right there tells you that he probably has quality stuff.

It’s different, probably, for prospect evaluation. At the lower levels, it might be more important to try to fill out repertoires, because that should only help against stiffer competition. But if a guy’s proven himself, he’s proven himself, no matter how he’s gone about doing it. By the time a guy’s in the majors, he’s passed some really hard tests. By the time a guy’s had early success in the majors, he’s only passed more.

Take a guy like Tony Cingrani. Cingrani would probably benefit from having a better change, or from having a more reliable breaking ball. Cingrani could absolutely improve — as a starter last year, he had an average FIP, and he didn’t work very deep. But then, literally everyone could improve, and Cingrani also struck out 28% of batters. He shredded his way through the minors. He’s already effective, and there’s little reason to believe hitters will suddenly be able to pound him after they adjust. Those adjustment periods are probably overstated.

There’s nothing more important for a pitcher than being able to get big-league hitters out. That is, there’s nothing more important than results. The process is important in its own right, but it shouldn’t be over-weighted. Worry about pitchers when there’s reason to worry. Don’t try too hard to anticipate those reasons. We’re not very good at it.




Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


34 Responses to “When a Limited Repertoire’s a Good-Enough Repertoire”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. R. A. Dickey says:

    I can tell when I’m not wanted.

    +45 Vote -1 Vote +1

  2. rusty says:

    I feel like the key step in the development of a 1- or 2-pitch starter happens before they hit MLB, and that’s not going to get picked up in this data set — when and how do teams give up on a guy and move him to the bullpen? The fact that limited-repertoire starters accumulate WAR at the same rate as 3+-pitch starters just seems to say that teams are making the right decision most of the time.

    I will add, though, that I was shocked — shocked! — at the prevalence of limited-repertoire starters; more than a quarter of total innings is much higher than I would have expected.

    Vote -1 Vote +1

    • Iron says:

      Well some might be caught by having too wide of a repertoire, no?

      Say a pitcher throws 50% fastball, 30% changeup, 8% slider, 8% curve, 6% sinker. He’d be a two-pitch pitcher in this measure, but not really. I still don’t have a problem with Jeff’s rough 10% cutoff here, but maybe a better way would be to say a pitcher who throws >90% of just two pitches. That may reduce the number of ptchers by a bit. Or maybe I misread it and that’s what he did.

      Vote -1 Vote +1

  3. DNA+ says:

    At least late in his career, Randy Johnson threw quite a few splitters to go along with the fastballs and sliders. Not sure when he developed the splitter, but I guess it just shows that even with two all time great pitches, a third look might not hurt.

    Vote -1 Vote +1

  4. Richie says:

    Great article. Umm, but don’t you normally chat on Tuesdays??

    Vote -1 Vote +1

  5. paperlions says:

    Much of this, I think, is a semantical misunderstanding. An 0-2 curveball in the dirt is a different pitch than an 0-0 get it over curve. An 0-2 rising FB is a different pitch than a FB at the knees. In addition, many pitchers can add or subtract a little from the “same” pitch to get different action.

    These things are just as much a part of pitch sequence and having different ways to retire hitters as having differently named pitches.

    Vote -1 Vote +1

    • Bip says:

      Those pitches are the same pitches in at least a few meaningful ways. The 0-0 curve and the 0-2 curve in the dirt have the same movement and velocity. Common sense suggests that if a pitcher has two pitches – where a “pitch” is defined as a ball thrown with a particular movement and velocity – then hitters can guess which is coming more easily, and that they could more easily learn to track them. This would make it easier for a hitter to extrapolate where the ball is going and make contact or lay off accordingly.

      Obviously location is extremely important, but one can still argue why having only two pitches would be a disadvantage.

      Vote -1 Vote +1

      • paperlions says:

        Common sense wouldn’t say that at all. A curve in the dirt leaves the hand at a very different angle than a curve in the middle of the zone. The hitters have to react before they see movement. Velocity and angle of release or the spin of the ball (if they can see it) are all a hitter will have in order to predict pitch location/movement. In addition, hitters guess. A LOT. Being able to use the same pitch in different ways or locations is enough to get a hitter out. You don’t have to fool the hitter completely, you only have to fool him in one dimension or get him to hesitate.

        The point is that there are several ways to get a hitter out with the same pitch, there is not one way to get a hitter out with each pitch….and the amount of variation (location, speed, arm angle, etc.) in a “pitch” is important…which is why some guys can throw the same pitch all day and still fool hitters the 3rd time through the order.

        Vote -1 Vote +1

        • The Hoot says:

          Bob Gibson talks about your point in his autobiography, “Stranger to the Game”. He wrote that people consider him a two-pitch pitcher – just a fastball and slider guy. However, he used the slider many different ways to achieve different effects. He threw hard sliders, slower ones with more bend, backdoor sliders, etc. to give multiple looks. So he considered himself to have a more diverse arsenal than just two pitches. I agree with you that there is a more nuanced way to look at what a guy throws than the types of pitches he has.

          Vote -1 Vote +1

        • Bip says:

          My point isn’t contrary to anything you just said. But if a pitcher throws the ball at a particular angle, a batter still has fewer types of movement to try to discern if a pitcher only has two pitches compared to three or more. Your point that hitters guess is totally consistent with my point. My point is that if a pitcher has fewer pitches, guessing is easier in at least one respect.

          For example, if a pitcher has only a 92 mph fastball and a 78 mph curve, it will probably be easier to time his pitches than if he also has a 85 mph change.

          My point isn’t at all that a pitcher cannot be successful with two pitches, or that a pitcher with two pitches is likely to be worse than one with more. I might assume that given two equally successful pitchers, one with two pitches and one with three, that the one with two pitches is better than the other in some way, either because his two pitches are better than the other three, or that he has better command of them.

          Vote -1 Vote +1

  6. derp says:

    I’m surprised no AJ Burnett.

    Vote -1 Vote +1

    • Bip says:

      Or Jose Fernandez. Perhaps Fernandez was so good last year that no one is actually worried about him, despite the fact he basically has only a fastball and a curveball.

      Vote -1 Vote +1

      • Chad says:

        Have you seen his curveball? Best pitch in the Majors.

        Vote -1 Vote +1

        • Bip says:

          I have. According to some, having only two pitches is basically unsustainable for a starter. While that is clearly not the case, it does seem to me that in order to be successful as a starter with only two pitches, those pitches had better be damn good. Obviously that is not a problem for Fernandez.

          Vote -1 Vote +1

        • A Commenter says:

          Aroldis Chapman’s slider. Or his fastball.

          Vote -1 Vote +1

        • Benjamin says:

          Jose’s fastball is pretty damn good too…i remember being just dumbfounded during his debut :s

          Vote -1 Vote +1

      • stew says:

        He also has a really good change. He’s not lacking for a third pitch.

        Vote -1 Vote +1

    • Burnett is also a very good example. And Fernandez, although BIS didn’t pick up on that.

      Vote -1 Vote +1

      • Bip says:

        I think a pitcher with multiple fastballs may not be included here. Fernandez has a four-seamer and a two-seamer according to pitch fx. I don’t know if that should exclude him from the criteria of this article though. The two fastballs are the same speed.

        Vote -1 Vote +1

  7. BDF says:

    I think of Dwight Gooden as the ultimate successful two-pitch pitcher, and some accounts tie his decline (only in part!) to his development of a changeup of which he became too enamored.

    Vote -1 Vote +1

  8. Bip says:

    Kershaw himself was a limited repertoire pitcher in 2010 and 2011, when he put his curveball on the back-burner due to control issues, relying on his slider as his only offspeed pitch. I’d say those two seasons were pretty pretty ok.

    Vote -1 Vote +1

  9. mockcarr says:

    Then there’s a guy like Detwiler who doesn’t really have a breaking ball or change he’ll throw even 10 percent of the time. It’s really sad, because when he came up his curve was probably his best pitch, but he lost that at some point.

    Vote -1 Vote +1

  10. Dirtbags says:

    I know it’s not the main point of the article, but I just want to point out Webb was not a two-pitch pitcher. His changeup was considered one of the best in the league, and he used it 33.8% of the time in 2009 all the while still using his sinker and curve (15.6%).

    Vote -1 Vote +1

    • Webb’s 2006 made the cutoff. If that’s inaccurate, blame BIS! Don’t blame Sullivan!

      Vote -1 Vote +1

      • Dirtbags says:

        If they are looking at early Webb, that would make sense.

        I remember an interview where Miguel Montero said he was catching Webb who was fooling around with grips, and all of a sudden he threw the best changeup he had ever seen. Montero convinced Webb to make it a bigger part of his repertoire. This was later in his career when his sinker was losing some velocity, so it fits in with the “evolve when batters force you to” mantra.

        Vote -1 Vote +1

  11. Chad says:

    Interesting article. I’m wondering about James Paxton and his chances of success next year. He come up last year and dominated, but had only the fastball and curveball; I’m not sure if he’ll need another pitch.

    Vote -1 Vote +1

  12. Mr Punch says:

    Seems to me the fastball-curve parlay isn’t great in the long run – once the fastball slows, it’s good to have another pitch off the same motion. Josh Beckett had a third pitch in ’07, but suffered when he lost it; Brad Penny struggled even retaining his velocity; Ben Sheets is hardly a counter
    example.

    Vote -1 Vote +1

    • Bip says:

      That’s a pretty anecdotal sample. The curve is the slowest pitch, so one might hypothesize that it would be the most effective with declining fastball velocity, considering it would maintain a large timing difference with a fastball.

      As for examples of guys with long term success with only a curveball, I can think of Gio Gonzalez, who is still doing quite well. If you count a cutter as a fastball, then there are some very encouraging examples of players who basically have only a curveball and various fastballs, like Wainwright and Halladay (before he developed his split-change).

      Vote -1 Vote +1

  13. stew says:

    A third pitch might be necessary if neither are plus-plus, 70 or better pitches but otherwise, I think the axiom is bunk. Curt Schilling was one of the most extreme 2 pitch pitchers of all time but it didn’t hurt him because his fastball and splitter were dominant pitches. It was ridiculous how repetitive his sequences were: fastball away, fastball away, get a foul ball or called strike and then a splitter. I don’t think I saw him throw a pitch besides a fastball or splitter until after he hurt his ankle and his stuff diminished, making it necessary to give hitters another look even if from a terrible curveball.

    Vote -1 Vote +1

    • David Berg says:

      Every dominant strikeout pitcher I can think of was a 2-pitch pitcher at his peak. Kerry Wood, Nolan Ryan, Sandy Koufax, etc. Curt Schilling and Roger Clemens actually had MORE pitches early in their careers; over time, their splitters simply became so much better than their other secondary pitches that they pared their repertoires down.

      With change-ups, splitters and breaking balls being largely “feel” pitches, it makes sense to me that it’s easier to throw a really good one if you throw it over and over, as opposed to busting it out 10 times a game. Yeah, you lose the element of surprise, but that’s probably worth it if you locate something with a nasty break on it.

      For an example of what NOT to do, see Mike Pelfrey. He was drafted with a great sinker, decent curve, mediocre change-up. In the minors, he scrapped the curve and added a slider. He had his first good year with this repertoire in 2008. After a bad start in 2009, he added the curve back and started throwing more 4-seamers. In 2010, he replaced his change-up with a splitter. In 2011, he added a cutter. He never learned to consistently command ANY of these pitches.

      My pitching philosophy would be to find the ideal secondary pitch for a given pitcher and focus on that. Just think if Randy Johnson had continued trying to use his curveball instead of committing to the slider.

      Vote -1 Vote +1

  14. randplaty says:

    Tyson Ross is another two pitch guy. Nasty slider though.

    Vote -1 Vote +1

  15. Baltic Fox Has Cold Paws says:

    Jim Palmer and Doc Gooden. Both of them were essentially fastball/curveball pitchers. Of course, when you have exceptional command of a limited arsenal, you can get away with that.

    Jim Palmer would be pals with Jeff Sullivan. During the 2012 season, asked by his fellow TV commentator what suggestions he would give Jake Arrieta he said: “Stick to the pitches you can command.”

    He then went on to echo what Jeff is saying in this piece. If you can’t command your third and fourth offerings, save them for bullpen sessions. Getting ahead of batters with the pitches you can command is more important than demonstrating to batters that you have a wide arsenal–and falling behind them in the process.

    Palmer said he tried for years to master the changeup, but he just couldn’t do it. So he rarely used it. He did mix in a slider, but he strictly limited how often he used it because he was convinced that over-reliance on it would lead to arm problems. (TJ surgery came around late in Palmer’s career.)

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>