Halladay’s New Changeup

Yesterday I read that Roy Halladay had changed the grip on his changeup and was throwing it more often (hat tip to Calcaterra on this one). Before this year Halladay held the ball in his palm when throwing a changeup, but during the offseason he worked with pitching coach Rich Dubee and changed to using a split-finger grip. Let’s see whether we can pick up the difference in the pitchf/x data. I think it is clearest in velocity-vertical spin deflection (vertical movement) space.

Definitely a difference: the changeup has about the same velocity but instead of rising about five inches it now drops a couple of inches. The horizontal movement on the pitch is also different.

So the different grip imparts different movement on the pitch, so I think it is fair to say it is a qualitatively different pitch. And Halladay is throwing the pitch more often. As is noted in the article, before 2010 Halladay was pretty much a three-pitch pitcher (cutter, sinker and curve — three great pitches), rarely throwing his change more than 5% of the time. But now he is throwing the change almost 12% of the time. It seems he is much more comfortable with it.

The results show that Halladay has good reason to be more comfortable with the pitch. By our linear weights it is worth 1.5 runs per 100, much better than his changeups in any full year before. Looking at components, the pitch is wildly more successful, getting 19% swinging strikes per pitch in 2010 compared to 6% in 2009-2007 (pitchf/x years). It also gets more ground balls (57% GB/BIP versus 55%) and a lower slugging on contact ( 0.452 bases/contact versus 0.505).

Looking at Halladay’s pitch-count splits, he uses the pitch often when he is ahead in the count. So it looks like he has developed a second out pitch — along with his curve — to put batters away when he is ahead in the count. As Calcaterra noted, it is not like Halladay needed another weapon, but he has one.

Print This Post

Dave Allen's other baseball work can be found at Baseball Analysts.

40 Responses to “Halladay’s New Changeup”

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

    Wow, how scary is that, one more above average pitch, in case he wasn’t already dominant enough.

    Vote -1 Vote +1

  2. PhilliesBullpen says:

    This makes me happy to be a Phillies fan because the idea of facing Halladay is just scary.

    Vote -1 Vote +1

  3. tdotsports1 says:

    As an avid Jays fan and self proclaimed Halladay lover, I definitely noticed the increased usage of the change-up during his starts for the Phils (thank you MLB.tv).

    Shaun Marcum (Mr. Change-up) and Halladay apparantly had a long running joke about Halladay’s change over the years, specifically his lack of one. Now it is amongst the leaders in run values/100 pitches, well it’s up there anyway.

    Vote -1 Vote +1

  4. Mike Green says:

    Excellent analysis, Dave. I have maintained for years that if there is one pitcher in mid-career who I expect to be great at age 40, it would be Halladay.

    Vote -1 Vote +1

  5. Ben says:

    I’m still fairly fresh to the SABR world, especially the more complex breakdowns such as this, so I could just be ignorant here. But does this take into account at all the quality of hitters that he’s facing? I’m speaking specifically about the success rate of the change compared to years past…unless it’s already taken into account (which is entirely possible), wouldn’t we assume at least some of the difference comes from facing less potent hitters than exist in the AL (especially the East) and facing the pitcher 2-3 times a game?

    Vote -1 Vote +1

    • Steve says:

      You are right, but the point was mainly that his change is moving differently this year than in years past. And he’s throwing it more often. Its a completely different pitch.

      Vote -1 Vote +1

      • Ben says:

        Yeah, that’s well taken and it’s a good point. But if we’re going to explore his new pitch, it seems necessary to try to evaluate its effectiveness. Otherwise it’s not particularly relevant that he’s throwing a different pitch. But I’m asking as a curiosity, not a criticism.

        Vote -1 Vote +1

    • don says:

      The effectiveness of his changeup has improved relative to his other pitches which should account for the different quality of hitter (unless the AL East hitters were for some reason particularly good at hitting changeups relative to other pitches, but that seems unlikely). I think the pitch type linear weights are a fairly ‘noisy’ stat, though.

      Vote -1 Vote +1

      • Dave Allen says:

        Ben, good point about the quality of the hitter. This analysis does not correct for it. Though I think that the fact that Halladay throws it more often is an endorsement that it is a better pitch.

        Don, yeah the linear weights are very noisy and very luck dependent (as well as sequencing dependent) so they should not be taken as gospel. That is why I included swinging strike rate and GB% which should be a little less noisy and luck dependent.

        Vote -1 Vote +1

  6. don says:

    Interesting that the ‘new’ changeup is not the one thrown by Moyer or Hamels, who would seem to be guys who know a thing or two about changeups.

    I wonder if anyone throws both.

    Vote -1 Vote +1

  7. Josh says:

    I’m patiently waiting for Halladay to develop a devastating slider. That’s pretty much the only pitch he has left to learn.

    Vote -1 Vote +1

    • fjkagreklg says:

      With his cutter and curve he really doesnt need a slider.

      Vote -1 Vote +1

    • BS says:

      Don’t forget about The Halladay Knuckler. =P

      Vote -1 Vote +1

    • 198d says:

      Halladay can throw the slider and the splitter, actually, and used them both a bit early in his career. However, IIRC, he found each to be fatiguing and developed the cutter instead. Seems like a great decision, given his results.

      Vote -1 Vote +1

  8. Matt Lentzner says:

    Nice article, Dave. I can’t see how the pitch can get negative lift unless it’s actually rotating so slowly that there’s knuckling action going on. That would be *very* interesting, if true.

    I’d call that a forkball and not a splitter. This is my opinion, but a splitter is thrown hard with some backspin and is generally faster (-5 mph from the fastball) than a slider (typically -7 mph). A forkball is a true changeup pitch where the ball is held deep between the fingers. It has almost no spin and is thrown at changeup speeds (-9 mph or less). I’d like to see that distinction made since I feel they are two separate pitches.

    Vote -1 Vote +1

  9. Brady says:

    Based on this chart, am I to understand that Roy Halladay’s sinker actually rises ~5 inches? *brain explodes*

    Vote -1 Vote +1

    • 198d says:

      Sinker should probably read fastball given the velocity? Doc’s a FB/CT/CB/CH pitcher, and the other three are listed?

      Vote -1 Vote +1

      • Dave Allen says:

        Yeah most people use the terms two-seam fastball and sinker interchangeably. For most pitchers a sinker ‘sinks’ only in comparison to a four-seam fastball, which rises more. Guys with a low arm angle (Ziegler, Masterson) get real ‘sink’ or negative vertical spin deflection on their two-seam fastballs.

        Vote -1 Vote +1

  10. NEPP says:

    You don’t realize how dominant Doc is until you watch him pitch every 5th day. He just gets it done every time. Even when he doesn’t have his stuff, he still battles. Its very impressive to watch.

    Vote -1 Vote +1

  11. CircleChange11 says:

    Does he split it in between his 1st 2 fingers (index and middle) or split it between middle and ring finger (aka, Vulcan Change).

    Vote -1 Vote +1

    • 198d says:

      Looks like index and middle. Mike Fast posted a pretty sweet photo above. Here’s the re-post, credit to him.


      Vote -1 Vote +1

      • CircleChange11 says:

        That’s interesting. We have a pitcher that throws from the “side”, would actually be 3/4’s … like Halladay. As such he doesn’t have the same effectiveness on a curve or standard change. So, we have him throw a cutter, slider, and for a change he throws a split (same grip as RH34).

        I’ve also seen splits that go down and in, like a sinker.

        Personally, I wonder if it has to do with hand size. Roy looks like one of those guys that has huge hands. If your hands are too big, you can wrap them around the change too much, or you end up holding it too much on the fingertips … but that same disadvantage with a change, it a major plus on a split.

        Vote -1 Vote +1

  12. John says:

    Dave, am I reading correctly that you said his changeup “rose” last year? Even if that’s an average, or if it was a bad pitch, I can’t see a guy as talented as Halladay throwing a changeup that goes up.

    Vote -1 Vote +1

    • Shazbot says:

      That’s a relative thing. A sinker has ‘rise’ compared to a neutral pitch, but less than a normal fastball. A changeup that rises less than a normal fastball wouldn’t be considered one that goes up strongly.

      Vote -1 Vote +1

      • Mike Fast says:

        Right, Shazbot. Dave is talking about the pitch considering effect of spin and neglecting the effect of gravity. That’s a useful way to figure out what a pitcher is doing to the baseball. Gravity obviously makes a pitch drop a lot, and a slower pitch like a changeup has even more time to drop due to gravity than a fastball does. A changeup will end up dropping a couple feet, a fastball maybe a foot.

        Vote -1 Vote +1

      • Dave Allen says:

        Great point Mike. I did not make that clear, thanks for bringing that up.

        Vote -1 Vote +1

  13. BK says:

    It seems to me that there is a bigger gap between his changeup and cutter/sinker velocity, which might also be part of why it’s been such a good pitch.

    Vote -1 Vote +1

  14. Steve C says:

    On the results side of things he is doing quite well against righties this season. Higher k/9 and lower xFIP.

    Vote -1 Vote +1

  15. John says:

    Clearly I’m far behind in the conversation, but a sinker doesn’t sink, it rises?

    Vote -1 Vote +1

    • Steve C says:

      For most pitchers this is correct, it just rises less than a typical four seam fastball. As the four seamer is far and away that most common pitch in baseball a sinker does sink relative to it. On X and Z plots, the origin is a pitch thrown with no spin, and relative to this it rises.

      If a pitcher drops their arm angle enough the pitch will begin to actually sink and this is a pretty good way to get your GB% up over 60%.

      Vote -1 Vote +1

      • Wilsonian Democracy says:

        The word “rise” is also a tricky one in this context – no pitch “rises” insofar as every ball thrown succumbs to gravitational forces and gets yanked down towards the ground between the mound and the plate. The “rise” here is relative to anticipated trajectory. Certain pitches maintain their plane against gravitational forces longer than others. A two-seamer thrown from a 3/4 arm angle like Doc’s gets backspin cutting across the vertical and horizontal planes at about a 45 degree angle, and this accounts for its fading motion. This can have the result of appearing to the batter’s eye as if the ball is “rising” from the trajectory it was on. In reality, the ball is darting down and in on a RHB, and a typical righthanded swing plane that misjudges the ball’s movement will end up on top of the ball and rolling it over to the left side of the infield.

        Vote -1 Vote +1

  16. CircleChange11 says:

    Guys, these pictures are amazing … and show exactly why Halladay is so durable and efficient.

    Balance and Staying Closed


    This pic is beautiful.
    [1] Balanced.
    [2] Front shoulder is closed.
    [3] Front knee goes toward back shoulder (closed).
    [4] Front knees splits the hands.

    He is able to get “coiled” (stay closed) or create torque without turning his back to the catcher (over-rotation).

    Stride Length & Hand Seperation


    Most pitchers have a stride length of 80% of their height. The goal should be (IMO) 100% of height. Tim Lincecum is 120%. Most 9-14yo’s are about 50%, and it drives me crazy. I find myself wanting to go to the mound and help the other team’s pitcher. Halladay’s stride length by my estimation is right about 100%, maybe a touch more.

    He also takes the ball back as far as possible, without doing so in an extreme way (Rich harden).

    Tall guy, long stride, shorter distance to home plate for the ball.

    Front Foot Landing & Glove Position

    When front foot lands, the throwing forearm is vertical. The front glove is tucked and thefront elbow is ready to come in contact with the front hip.

    I REALLY love that the front elbow is “higher” than the back elbow. IMO, a high front shoulder leads to a lower back shoulder (good), and the only way to have both shoulders “high” is to “Prior It” and do the inverted W.


    Easier on the arm. More movement on pitches. Eaiser to keep the head “stable” and balanced (nose over front knee). Glove side arm tucked closely to the body (awesome, complete opposite of injury guys like harden and calero). Skinny guy has some muscular legs. The arm slot has to be murder for RHB … especially when he (foot, glove, arm) sort of “sweeps” out of balance point, toward the hitter with a long stride.

    Halladay is actually similar to Pedro. Pedro was lighter and more violent in his motion, but very similar in balance position, stride, arm slot.

    Vote -1 Vote +1