CC Sabathia's Opening Day start through the lens of PitchFX

(photo c/o The AP)

When we last saw CC Sabathia in action prior to Thursday’s Opening Day start (and subsequent Yankee win), he was busy gutting his way through a six-inning, two-run performance against the Rangers in Game 5 of the 2010 ALCS to keep the Yankees’ season alive.

Interestingly, Thursday’s start was actually not all that dissimilar from that ALCS Game 5 performance — in both outings he went six innings and gave up two earned runs, and, if you take a look at the chart in the aforelinked post and compare it to the table below, his average velocity and H- and V-breaks on all five of his pitches in Thursday afternoon’s game were actually pretty close to the numbers he registered in that last October start.

CCs best pitch on Thursday was his sinker, which he used about a quarter of the time and registered -2.02 linear weights for (for those that need a refresher, negative linear weights for a pitcher’s pitches are a good thing — the more negative, the better). His change was his least effective pitch, though he only went to it about 10% of the time.

Sabathia got 14 swinging strikes, which is wonderful, though of course a good portion of those were committed by Austin Jackson and Alex Avila, who looked pretty lost at the plate against CC.

Unsurprisingly, his two highest-pitch innings were his least effective, but on the whole those are some solid linear weight tallies on an individual inning basis. Sabathia probably could’ve gone back out for the seventh, but why on earth do that in the first game of the season unless you absolutely have to?

Below are CC’s average pitch breaks for the entire 2010 season, courtesy of

On Opening Day he threw his fastball further inside (to lefties) and finished more than two inches higher on average than he did in 2010. He also threw his sinker nearly two inches further inside (again, to lefties) and the pitch finished more than four inches higher on average than it did last season. Given Sabathia’s success with the pitch on Thursday, I wonder if we’ll be seeing more extreme movement on the pitch in 2011. The change also broke further inside and higher. The slider had slightly less horizontal break on righties than his average 2010 slider did, and, like all his other pitches, broke higher than last season. And his curve on Thursday was further inside to righties than on average last season, but broke a good deal less (-0.78 V-break compared to -3.35).

So Sabathia finished all his pitches higher than usual except for his curve — given that this was his first outing of the season, this is not terribly surprising.

Here’s Sabathia’s strikezone plot, also courtesy of Brooks:

Nothing too outrageous here; looks like Sabath received just one gift strike call, though he may have missed out on a few low called strikes.

All in all, a typically strong performance from the big man, and I’ll be looking forward to him building on this and getting even better as the season progresses.

13 thoughts on “CC Sabathia's Opening Day start through the lens of PitchFX

  1. I was convinced he threw the curve on the first pitch of AB’s at least 6 or 7 times. I may have made an error and mistaken a slider or sinker for a curve, but I wasn’t the only one. Interesting.

  2. You may have been right — the one frustrating aspect of working with the PitchFX data is that the pitch classifications are never 100% accurate. It’s entirely possible some of those curves got mistaken for sliders.

    In fact, Gameday had Sabathia starting out with a curve against Ordonez in the 1st, V-Mart in the 2nd, Inge in the 2nd, Jackson in the 3rd, Cabrera in the 4th, Inge in the 4th, Ordonez in the 5th, Martinez in the 5th, and also throwing a curve for the 2nd pitch of the at-bat against Inge in the 6th. All told, that’s actually 8 batters that saw curveballs on the 1st pitch of their a-bats (according to Gameday), and nine curveballs total, so it looks as though your eyes did not in fact deceive you.

    This is a head-scratcher, as Brooks clearly states that “Pitch classifications provided by the Gameday Algorithm” though of course it follows that up with “and may be inaccurate.”

    Sabathia got a strike on eight of nine of those Gameday curveballs, and none were turned on for hits, so perhaps some of the linear weights attributed to the slider by Brooks should actually have gone to the curve.

    I imagine this will probably be a season-long problem, and I’ll be sure to be mindful of the fact that for CC, some curves may be mis-classified as sliders when analyzing the PFX data.

    Maybe one day we’ll have a flawless pitch classification system.

    • Why don’t you contact Brooks? There is a form or a forum or somethign somewhere on that site, and he appreciates when mistakes are pointed out.

      • I started doing PitchFX profiles of every starting pitcher’s game in the middle of last season, but scrapped it because the data was too unreliable. Half the time, the pitch count would be off by as many as 3 or 4, and when I made notes of particular pitches, they often differed from what was reported by Brooks.

        • It’s definitely frustrating because I greatly enjoy parsing the data.

          Hopefully once the season gets underway the data will become somewhat more reliable; I’d hate to have to scotch the pitchfx analyses entirely.

  3. The velocity will tell you if CC was throwing a curve or a slider. The chatter on twitter yesterday was that his slider is significantly harder than his curveball. Not 100% on that, but I believe that’s what I had read.

    • Thanks Michael — I just went back and checked, and the nine pitches I identified as curveballs in the above comment were definitely all curveballs — low-70s velocity.

      I have an e-mail into Brooks asking if it’s possible to fix the data.

      • In the post game, CC specifically spoke about throwing his curve early in the count for strikes, so yu are absolutely correct.

  4. No problem, Larry. I find following Harry Pavlidis (@HarryPav), Josh Smolow (@garik16), Albert Lyu (@thinkbluecrew) and Lucas Apostoleris (@DBITLefty) if you don’t already. Mike Fast, too (@fastballs) as they all usually engage in PITCHf/x discussions on there and it’s awesome. I learn so much by lurking their back-and-forth’s on there, mostly about random pitchers they’re all looking at.

    • Awesome; I already followed Albert and Mike; just added the rest of those guys as well. I love PitchFX, and those guys are the cream of the crop, definitely looking forward to hearing what they have to say.

  5. Hi All-

    You can see on this graph here:

    What’s going on is the pitch class algorithm is incorrectly labeling two curves as sliders or vice versa.

    You can pull up those graphs by selecting “Pitch Type” when sorting.

    I don’t run the algorithm; it’s run by MLB. I just regraph the data. And, I just checked – all 106 pitches are scored in that file.

    So, sorry about the mislabeling =). It happens, but it’s growing more and more infrequent over the last few years.


  6. Pingback: Get Ready For Spring With Mlb Predictions | Wind And Solar Power For Your Home