First-Pitch Curveball: A Whodunit by Ben Clemens July 22, 2020 Think of the stereotypical curveball thrown to start an at-bat. Picture it in your mind’s eye. It’s big and loopy, starting high and then swooping into the zone to steal a strike from the incredulous batter. It’s an optical illusion, a strike disguised as a ball. It’s probably more or less 12-6 when it comes to break; a perfect rainbow from pitcher’s hand to catcher’s mitt. And if it’s an active player throwing it, you’re probably picturing Rich Hill. In some ways, you’re not wrong. Rich Hill does throw a ton of curveballs, and first-pitch curveballs are in the strike zone far more often than hooks thrown on every other count. Hill isn’t the foremost practitioner of the art, however. Of the 359 pitchers who faced 100 batters and threw at least one curveball last year, Hill had the 41st-highest first-pitch curveball rate at 32.4%. He was just outside the top 10% of the league, not out front by a mile. In fact, relative to how often he throws his curve, Hill is one of the least likely pitchers to throw it on the first pitch. On non-first-pitches, Hill threw it 44.4% of the time, 12 percentage points more often. Only 12 pitchers had a bigger negative differential when it came to starting batters off with curveballs relative to the rest of their pitch mix. Okay, so a first-pitch curve is kind of, but not really, a Rich Hill thing. His are mainly an artifact of his overall curveball mastery. Who, then, should your mind’s eye picture when we’re talking first pitches? That’s right — Austin Brice. Nearly 400 qualified pitchers, and no one threw a first-pitch curveball more often than Brice. Maybe you wouldn’t even consider that a curveball, and that’s fair. There are classification issues to consider — our two sources for pitch-level data disagree on whether Brice throws a curve or a slider. Whatever you call it, he was hucking it over the plate on 65% of first pitches last year. The top 10 of the list makes decent sense; you’ve got Nick Anderson on there, and Brandon Workman as well. Secondary-heavy relievers throw a lot of breaking balls on the first pitch because they throw a lot of breaking balls. When you have a pitch like Anderson’s, you might as well use it as often as possible. Ah, delightful: Extra points if you can start it in the middle of the zone, end it at the bottom of the zone, and still make it unhittable. Sorry, Aaron: But I can’t just keep showing you clips of Nick Anderson embarrassing batters, as much as I’d like to. We were talking about something here! First-pitch curveballs, and the wackos who lean on them. Gotta keep moving. If we want to look for pitchers who disproportionately lean on first-pitch curveballs, we can’t just look for pitchers who throw the pitch often. We have to look for pitchers who turn to their curve on 0-0, the reverse Rich Hills of the world. The top of that list is altogether more confusing: Disproportionate Curve Users Player 0-0 Rate Overall Rate Difference Jace Fry 44.2% 10.7% 33.5% Yimi García 52.6% 23.6% 29.0% Evan Marshall 37.9% 9.7% 28.2% Austin Brice 65.0% 37.8% 27.2% John Gant 30.6% 3.7% 26.9% Adam Morgan 32.2% 7.0% 25.2% Derek Law 38.9% 13.9% 25.0% Junior Guerra 36.3% 12.4% 23.9% Blake Parker 30.7% 10.4% 20.3% Noé Ramirez 51.4% 32.0% 19.4% What in the world is Jace Fry doing up there? He’s nearly half curveballs on the first pitch of an at-bat, then only 10% the rest of the time. How different is 0-0 from 1-0 and 0-1, when he threw 19% curves? For that matter, what about John Gant? A third curveballs when he normally throws the pitch only three out of 100 times — that’s nonsense. Michael Wacha, the pitcher who first clued me into this behavior, is just off the list in 13th; he starts with 23.8% curves and throws it only 5.7% of the time otherwise. Why do pitchers who, in the main, don’t trust their curve use it so often to start an at-bat? I think I know the answer. But like a filmmaker cleverly stringing his audience along, or a writer padding word count, I’ll merely nudge you in the correct direction rather than screaming it. Here’s an interesting fact about the 10 pitchers who increase their curveball usage most on 0-0: their curveballs are bad! They have an average run value, per 100 curveballs thrown, of -0.25. The top 30 checks in at -0.5 on average. The bottom of the scale, on the other hand, is comparatively excellent. The 10 pitchers who decrease curve usage most on 0-0 have an average run value of +1.1 runs per 100 pitches. The bottom 30 are at +0.8. Here’s a misleading, but still honest, way of putting it. 0-0 is the most consequential pitch of an at-bat, a pitch thrown every time and one that has huge bearing on how the rest of that at-bat goes. The pitchers with the worst curveballs love nothing more than to throw that pitch on 0-0. The pitchers with the best ones? They never use it! They use their fastballs. Maybe you caught the deception there. I got close to it, if nothing else. You see, 0-0 might be an important pitch, but it’s also a very weird one relative to the counts that normally call for breaking balls. You know how you accrue value with a breaking ball? With whiffs, glorious piles of two-strike whiffs that build a run-value castle. The pitchers with the best breaking balls are, generally speaking, the ones with filthy wipeout pitches, the kind of offerings that can be deployed at will to trick batters defending the plate. You know what almost no batter does on 0-0? Defend the plate. Throw them a two-strike curve, and this is liable to happen: Better luck next time, thanks for playing. If you can’t get batters to swing, if your curve doesn’t quite have that snapdragon quality, you’re in a different bind altogether. If you can’t make batters swing at your curveball, then you can’t make them whiff at your curveball in the dirt, and so you might as well not throw it at all. But wait! What if there’s a loophole? Batters can recognize your curve, sure, and that’s your problem. Something in their brain tells them not to swing. If you threw it to them on 0-0, when they already don’t want to swing, they’d surely just automatically take, right? Then why not just throw it right in the strike zone? I exaggerate, but that’s a good general sketch of the thought process. Pitchers who know, absolutely know, that they won’t need their curve again the rest of the at-bat love throwing it for a strike on the first pitch. They don’t need to worry about the batter getting a good look at it, and if it’s taken for a strike, it was a good deal for the pitcher. Look at this! That’s not a curve you would try to spike into the dirt. Nobody was fooled there; Austin Slater saw spin and decided to take, and Gant said thank you very much. That would be a great plan — if hitters never swung. The only problem is, hitters started swinging. Take a look at a mouthful of a statistic — swing rate at in-zone curveballs thrown on a 0-0 count: Declining Automatic Takes Year Swing Rate 2010 29.4% 2011 30.1% 2012 29.1% 2013 30.0% 2014 30.2% 2015 31.8% 2016 31.7% 2017 33.7% 2018 34.3% 2019 35.7% In other words, that free strike is becoming less free. The pitchers who are leaning most on first-pitch curves allow louder contact, worsening the blow. Those top 30 0-0 curve users from before? They allowed a .390 xwOBA on contact in 2019. The bottom 30 were 60 points better at .330. You might argue there’s some confounding variable there. The pitchers hanging the most curves on 0-0 have the worst numbers on contact because batters get to pound on pitches in the strike zone. Maybe that’s true! But if that’s the case, why hang pitches in the strike zone in the first place? It’s more complex than I make it out to be; first pitches aren’t completely disconnected from the rest of the at-bat, not every pitcher who uses their curve a lot on 0-0 throws a bad one, and batters aren’t indiscriminately swinging at every single curve more often but are rather picking on weaker show-me offerings. As a general statement though, it’s not wrong to say that first-pitch curveballs, at least the type that pitchers rarely use in other counts, aren’t where you want to be. I’m not sure there’s much lesson to any of this, or anything that can be learned for 2020. Maybe pitchers with middling breaking stuff should stop throwing it as much to start at-bats, though pitching coaches surely already know that. Maybe batters should hunt curves even more from some pitchers. Mostly, however, all I’m telling you is that pitchers with bad curves shouldn’t throw them so much. In the end, the picture I asked you to imagine at the top of the article was a lie. The stereotypical 0-0 curveball isn’t thrown by Rich Hill, because the stereotypical 0-0 curveball isn’t good.