It’s a simple thing to say, but there’s an important interplay between the swing and the miss when it comes to pitching. In order to get a swinging strike, you need to get the batter to swing and you need to get them to miss. These are, in effect, two different skills, even if the best pitchers are awesome at both. And so it’s not surprising that we have two different metrics for that moment — whiffs per swing (whiff% in some places) and whiffs per pitch or swinging strike rate (swsTR% here). We probably need both. Is one better?
Let’s first have some fun and look at the outliers who are good at one of the interwoven skills and not at the other. In order to find these pitchers, I set high benchmarks for the use of each pitch and then ranked the qualifying starters by whiffs per swing and swinging strike rate respectively. Then I looked for those that were ranked highly in whiffs per swing but poorly by swinging strike rate. These starters threw the pitch in question regularly last season, and got whiffs once the batter swung, but didn’t get batters to swing as often as other pitchers. Weird.
Matt Cain, Change-up
39.9% whiffs/swing, sixth of 73. 16.4% whiffs/pitch, 33rd of 73.
Maybe this is just a one-year blip. The average change-up gets just under 30% whiffs per swing and features a swinging strike rate around 15%, and for his career, Cain’s change-up is right there on both counts. But last year, batters only offered at his change-up 40% of the time (baseball swung about 50% of the time at a change last year). So his whiff/swing number looked big while his whiff/pitch number was just a tiny bit larger than usual. Guess what also happened this year. Cain’s change-up was a ball ten percent more often than it was in the past. If you throw a good pitch for whiffs, but can’t put it in the zone, this is what can happen to you. This might been part of Cain’s trouble in 2013, actually.
Jordan Zimmermann, Curveball
35.2% whiffs/swing, 12th of 69. 10.1% whiffs/pitch, 37th of 69.
The curveball is an interesting pitch because it gets offered at less than any other pitch in baseball. Perhaps the big bend at the top is easy to spot, and the pitch goes for balls more than most other pitches, but either way it’s a pitch batters love to take. Zimmermann gets about seven percent more whiffs than your average curveball once he gets batters to swing, but batters only swung 28% of the time at the pitch last year, a full 10% worse than league average. The result is a swinging strike rate that’s right about where curves sit (11% is average), but he got there in strange fashion. Like with Cain’s change-up, though, this effect goes away if you zoom out and look at his whole career. Unlike Cain’s change-up, however, the ball percentage on his curve only differed by a couple percentage points over his career number. Maybe the league is adjusting to him a bit.
Chris Sale, Slider
37.9% whiffs/swing, 17th of 59. 14.1% whiffs/pitch, 40th of 59.
Where Zimmermann and Cain might only be one-year blips, Sale’s been showing this effect his whole career. Once batters swing at his pitch, it’s plus-plus, elite, as good as it looks when you’re watching it live. But while batters usually swing at sliders close to 50% of the time, they swung at Sale’s slider only 36% of the time last year (40% career according to BrooksBaseball). And his ball rate has remained mostly unchanged for his career. It worked for him, but still. This is a one-man argument for keeping both stats around.
If you’re looking for one number to hang on a pitch, it might be up to your preference between whiff percentage and swinging strike rate. My preference is to use the number that encapsulates both parts of the process — swinging strike rate asks how often you get both a swing and a miss when you throw the pitch, and that seems most important to me. If you want a year-to-year correlation for the two, swinging strike comes in at .804 and whiffs/swing at .789, which is another argument for swinging strike (albeit a mellow one). And, last, the denominator — the biggest difference between the two metrics — favors whiffs/pitch. Per-pitch numbers see their samples grow quicker than per-swing metrics, so they might be ‘believable’ (stabilize) quicker.
In any case, it’s good to have both numbers around. After all, you have to get them to swing and to miss if you want a whiff.
Print This Post