Kirby Yates and the home run

In case you haven't noticed, Kirby Yates has struggled at the major league level this season. He's only pitched 9.1 innings out of the bullpen, as he spent the beginning of the year on the disabled list, but over that one game worth of work he's posted an ERA of 7.71. That comes on the back of having given up five home runs out of 14 fly balls.

In his only other major league appearance (36 innings in 2014), Yates did not have trouble with the long ball.

Before I get any further into this, let's all acknowledge why analyzing 9.1 innings is a futile exercise: It's a sample size that's almost entirely noise. According to work by Russel Carleton, changes for HR/FB stabilizes (half signal, half noise) at around 50 fly balls for players who have given up at least 500 fly balls in their overall career. Yates has given up 14 this year, and 47 in his career.

Still, I thought, "Why not look for something?"

Surely there's a reason; surely there's a pattern. I noted that over his minor league career, Yates had shown a walk rate of around 10%, but in his 9.1 innings in 2015 he'd only walked 2.8% of the batters he'd faced (1 out of 42). What if that's a pattern? What if guys who suddenly start walking fewer batters give up more home runs?

To investigate, I pulled from the FanGraphs leaderboards every pitcher who pitched 50 innings both in 2015 so far and in 2014, and calculated their change in walk rates. Then I calculated their change in HR/FB rates. This is what I got.

Don't spend too much time looking at that. It's meaningless. While the players who improved their walk rate did give up a few more home runs, the r-squared of .0198 lets you know that it's not significant (if the blob of dots didn't already do that for you).

Anyone who's played with pitching numbers before will know that reality is noisy and confusing, and that it doesn't care for our easy small-sample-size narratives, but it's good to check and remind ourselves of that every now and again.

So what is going on with Yates?

Well, his fastball is pedestrian, having average velocity (slightly above, but close enough to call average) at 93 mph for his career, and below average movement. He does throw four different pitches, though, and he's apparently able to keep hitters off the fastball. It's actually has an above-average whiff rate in the majors (right at 20%, per swing, per Brooks Baseball).

Just looking at this graph, it looks like he's gotten a wide range of motion on his changeup, which probably is indicative of a few poor deliveries, but he's thrown it so infrequently that there's not much to be said. As for the home runs, two have come off the fastball, two against the slider, and one against the changeup. Nothing to make of that, either.

Pretty much all I can say is that Kirby Yates is going to stop giving up home runs at this incredible rate, because no one gives up home runs at this rate. Also, data is noisy, and sample size is real. Aren't you glad you came for the analysis?