The Return of Cahill’s Curve

Warning: The content of this post breaks two major rules recently laid out by the FanGraphs writing team. It’s a post about stats from last week, and it validates the first prediction in Matt Klaassen’s recent post. You’ve been warned, now proceed at your own risk.

Despite posting an ERA of 2.97, and winning 18 games last season, Trevor Cahill entered 2011 as one of the major regression candidates due to some poor peripherals. One of the most puzzling aspects of Cahill’s young career, and one of the reasons analysts are calling for regression, has been Cahill’s low strikeout rate in the majors. Over his minor league career, Cahill struck out 9.9 hitters per game. In the majors, however, Cahill has managed to strike out only 5.12 batters per nine innings. One theory behind Cahill’s struggles in the majors deals with the use of his curve ball. Minor league scouting reports rated Cahill’s curve as a true wipe-out offering. Once he reached the big leagues, however, he stopped relying on his curve. Well, the curve, and the strikeouts, returned in Cahill’s first start of the 2011 season. Unfortunately, those two facts have little to do with each other.

Last season, Cahill relied on his curve 13.6% of the time according to his PitchFx page. Of the 105 pitches Cahill threw in his start last week (and again, it’s a tiny, tiny sample), Cahill threw 24 curve balls, good for 22.8% of his pitches thrown that night. While early scouting reports might suggest Cahill’s increased usage of his curve led to his strong performance in the strikeout department (8 in 4.2 innings pitched), a quick look at his start reveals that’s not exactly the case.

Of his eight strikeouts, only two of them came via the curve. Cahill was given the inside corner vs lefties most of the night, and capitalized by getting four called strikeouts with his sinker. The other two strikeouts came via Cahill’s changeup. So, if Cahill didn’t utilize the curve as his strikeout pitch, what effect did it have during the game?

The main culprit behind his high pitch count may have been defensive misplays that extended innings, but Cahill’s curve may have contributed to his lack of efficiency as well. Out of the 24 curves thrown by Cahill, 6 were balls, 8 were strikes, 7 were fouled off, 2 led to hits, and 1 led to an out in the field.

With two strikes, those numbers become: 1, 2, 3, 2 and 0. Using that data, we can say that 13 of Cahill’s 24 curves led to a successful result (8 strikes, 4 fouls led to strikes and didn’t extend at-bats, 1 out in the field). It’s the other side of the equation, however, that is somewhat concerning.

With two strikes, Cahill’s curve wasn’t as effective. While he managed to strike out two batters, he also allowed three batters to extend their at-bat, which led to Cahill throwing more pitches. Again, it’s only one game, but the Mariners were able to fight off Cahill’s curve and extend their at-bats. When your defense commits four errors behind you, those extended at-bats really start to exacerbate your pitch count.

It’s not all gloom and doom with Cahill’s curve, though. The fact that he was willing to use it often will probably help him fool hitters over the course of the season. In his first start of the season, however, his increased usage of the curve did not lead to his high strikeout rate. It actually allowed Mariners’ hitters to extend innings and make Cahill throw more pitches. That’s not to say the curve won’t end up fulfilling it’s destiny as Cahill’s strikeout pitch, we just shouldn’t assume there’s a correlation at first glance.




Print This Post



Chris is a blogger for CBSSports.com. He has also contributed to Sports on Earth, the 2013 Hard Ball Times Baseball Annual, ESPN, FanGraphs and RotoGraphs. He tries to be funny on twitter @Chris_Cwik.


11 Responses to “The Return of Cahill’s Curve”

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

    Ou broke the rulez1111!!!!1!1!shift+one!

    Vote -1 Vote +1

  2. PL says:

    So….he’s still a regression candidate?

    Vote -1 Vote +1

  3. ofMontreal says:

    You know, I watched that start, and Cahill had nasty stuff. Absolutely nothing straight and always around the plate. It made his numbers from last season make a lot of sense. He’s a serious weak contact pitcher, so predict regression at your own peril. Not that he will win 18 again, that’s something else entirely.

    Vote -1 Vote +1

  4. Kris says:

    So yah. This article was written last year as well. Except Cahill was actually throwing his curve more often.

    The PFX system is all fucked to hell right now (i have no clue why) but various stadiums are adding a boat-load of sink. Because it’s all screwed, a lot of Cahill’s sliders got tossed into the curve pile.

    But yah, in one start, Cahill threw roughly 7 percent more breaking pitches. I love speculation and I think your readers are smart enough to understand when a small sample size warning must be applied.

    Cahill’s a good pitcher. If he wants to rack up some strike-outs, he’ll do it on the back of his change-up and curveball/slider. Whether or not it’s worth upping the PC is debatable.

    Vote -1 Vote +1

    • danmerqury says:

      You’re right in that everyone has a bizarre amount of sink, but I don’t think it’s the PitchFX system. It’s Fangraphs. Everything’s displaying fine on Brooks Baseball.

      http://www.brooksbaseball.net/pfxVB/cache/movement.php-pitchSel=502239&game=gid_2011_04_01_seamlb_oakmlb_1&batterX=&innings=yyyyyyyyy&sp_type=1&s_type=.gif

      Vote -1 Vote +1

      • Albert Lyu says:

        Going to have to correct you there. FanGraphs gets its pitchfx numbers from MLBAM (like most other sites) at the end of a day’s games. Brooks is the main site out there that gets it after each at-bat. The vertical movement numbers had more to do with post-gameday calibration issues than it did with pitch-by-pitch numbers — that is, the end-of-day pitch logs were re-calibrated and re-classified incorrectly across baseball for whatever reason, but looked fine on live updates.

        And, for what it’s worth, Cory Schwartz of MLBAM/MLB.com has indicated, just now, that the pfx vertical movement numbers should be resolved and more ‘normal’ than they were the first few days of the season. So we won’t see Barry Zito curves with -20 inches of vertical movement anymore, hopefully.

        Vote -1 Vote +1

      • Kris says:

        Yah, I have been somewhat confused by everything because I’ve been getting numbers from Brooks and Fangraphs. I think when I looked at Kyle Drabek, his Brooks starts (last year and this) were just as strange.

        I always hear so much about Drabek’s sinker, but it’s always appeared to be pretty ‘meh’ in terms of “sink”.

        Thanks very much for clearing that up Albert. It appeared to be pretty random when I looked at the data. Some guys appeared to have the same stuff and other guys had different stuff. Maybe it was just a difference between looking at Brooks and Fangraphs…. who knows…

        Vote -1 Vote +1

  5. Htpp says:

    I didn’t see Cahill pitch at all last year, but after seeing his last start, I’d never peg him for a ~5 K/9 guy. His pitches were doing some crazy things.

    Vote -1 Vote +1

  6. AJS says:

    It’s hard to know whether allowing three batters to extend at-bats is good or bad without context. It doesn’t seem like much to me.

    And if the fouls were batters barely connecting and making weak attempts to stay alive in the at-bat, how does that show the curve wasn’t effective? If he polished those hitters off on the next pitch with a fastball, I’m not convinced having to throw one extra pitch is a bad thing.

    Vote -1 Vote +1

    • Kris says:

      Well, allowing a batter to extend an AB is obviously a bad thing. How bad? I certainly don’t know.

      We don’t know the outcome of either event, so we’ve basically got to assume equal probability and thus, +1 to the PC is a bad thing.

      However, the way the Athletics defense is playing tonight, anything is better than inducing contact, even if it’s terribly weak contact.

      Vote -1 Vote +1

  7. hecubot says:

    It’s worth noting that the curve Cahill is throwing in the majors is not the same curve he threw in the minors when he was racking up those strikeouts.

    In the minors he threw a spike or knuckle curve.

    In the majors he’s been working with a true, snap-the-wrist curve. So it may not have the same movement or effect.

    Vote -1 Vote +1