My Favorite Andrew Miller Fact

Most of what I do here is provide you with fun facts. Let’s be real — you already have a decent idea of which players are good and which players are bad. A healthy portion of my job, then, is to tell you what you already know, but in some new and different way. When it works, I think we all get to come away feeling smart! Hopefully it continues to work.

What I have for you here is an Andrew Miller fun fact. Not just a fun fact — my absolute favorite Andrew Miller fun fact, at least of the moment, at least as long as it’s factual. It’s not like you didn’t already know that Andrew Miller is good. We all came to terms with that years ago, and Miller hasn’t gotten any worse. He’s gotten better! Boiled down, this post is just “Andrew Miller is great at pitching.” But there’s this thing, see. He turns hitters to brain-dead mush.

You know what O-Swing% is. It’s a PITCHf/x stat. It measures the rate of swings at pitches out of the strike zone. Similarly, you know what Z-Swing% is. It measures the rate of swings at pitches in the strike zone. Pretty obviously, batters are more likely to swing at strikes than balls. This year, the league-average O-Swing% is 30%. The league-average Z-Swing% is 63%. These numbers have remained fairly stable over the years.

Focusing now on Miller — setting an embarrassingly low minimum of 10 innings pitched, Miller owns baseball’s highest O-Swing%. (He’s easily cleared that minimum, but I just wanted to be inclusive.) That’s a good thing for Miller, because pitchers love it when hitters chase. Now we can move a column over in the leaderboard. Miller also owns baseball’s lowest Z-Swing%. That’s another good thing for Miller, because a taken strike is a free strike. No damage possible on taken strikes!

That gets us most of the way to the fun fact. Let’s put it all together now. Miller doesn’t just lead the league in both statistics. As I write this on Wednesday afternoon, Andrew Miller actually has a higher O-Swing% than Z-Swing%.

Think for just one second about what that means. One second is all you should need. Batters have been more likely to swing against Andrew Miller when he’s thrown would-be balls. That’s an absurdity, something that shouldn’t be possible, and even though I know we’re working with a smaller sample size, look at the current top 10 in swing-rate difference:

Miller’s O-Swing% is two points higher than his Z-Swing%. Perez is second-best here, yet worse than Miller by 15 points. The difference between first and second place is the same as the difference between second and 77th place. Statistically speaking, Andrew Miller shouldn’t exist. And though this particular truth doesn’t hold up if you expand the sample to the last calendar year, Miller does remain in the big-league lead in this measure.

It’s pretty clear this isn’t a fluke — Miller simultaneously discourages swings, and encourages them. He just encourages hitters to take the wrong swings, and he’s doing that in 2016 more than ever. Some short time ago, I wrote about how Aaron Nola is good at this. He’s very good at this! Miller is just on another planet. I also remember writing, way back, about how Henderson Alvarez had a higher O-Swing% than Z-Swing% as a hitter. Andrew Miller has effectively turned hitters into bad-hitting pitchers.

This plot shows the course of Miller’s progress since the dawn of PITCHf/x:

Just in case you missed that:

You can see Miller getting better and better, and now the lines have crossed. I don’t know if they’ll stay crossed — it’s hard to imagine they could stay crossed — but Miller most assuredly has this skill. It’s the skill of strike-zone deception, and it appears to be related to increasing slider usage. Batters don’t know what to do with it, and Miller has a keen sense of when to throw a slider in the zone, or when to throw a fastball. Here is how Miller has used his pitches this season, with fastballs on the left and sliders on the right:

The slider works right off of the fastball, darting glove-side and down. Miller has been consistent with how he’s thrown it, controlling the pitch better than ever, which leads to a 62% rate of sliders thrown, which leads to the following plot of swing rates by area:

As a hitter, you go up expecting the slider. It’s Miller’s primary pitch, so it’s right there in a hitter’s mind. Because hitters are thinking about the slider, Miller can put a fastball by them — only R.A. Dickey has a lower fastball Z-Swing%. But the slider is still overwhelming, because Miller uses it with such expert precision. It still looks like a fastball until it does things fastballs don’t do, and that’s when batters are left helpless. Miller can think along with them, and when a pitcher has smarts, stuff, and accuracy, that’s how you get something like Miller’s present stat line.

Here is a representative plate appearance. It’s not always this easy for Miller, but it might as well be.

It’s not representative, in that Miller threw twice as many fastballs as sliders. But Jean Segura was not expecting that, which Miller used to his own advantage. The two fastballs were perfectly placed, and then the slider was perfectly placed, and even when Segura was probably expecting the slider, it didn’t matter for a hill of beans. A good pitch, thrown properly, is almost untouchable. Andrew Miller throws only good pitches.

What Miller has done to this point is almost genuinely unbelievable. I do believe it. I have to — I wrote a whole damn post about it. But even I wonder about some kind of glitch. Something that explains why Miller looks so extraordinary. What I have to conclude for now is that Miller is just a glitch of nature.