Jake Arrieta won the 2015 Cy Young by spinning in sliders/cutters/slutters with a wipeout curve and heavy sinker. His dominant season sharply contrasted last season, when he looked slightly lost at times, much to the consternation of even the normally unperturbed Joe Maddon.
“Yes, of course there is [concern],” Maddon said. “We have to figure out exactly what’s going wrong. Yes, I cannot deny that, and I’m sure he’ll say the same thing.”
What’s remarkable, though, is that despite almost doubling his walk rate, Arrieta still produced 3.8 fWAR and a 3.10 ERA.
So what was the difference between Cy Young Arrieta and 2016 Arrieta? While the answer to this question is certainly multifaceted, I’m pretty sure I found one of the factors: in-game release point consistency of his slider. I ran a few tests on his release point consistency, the results of which are extremely convincing.
The first part of this post will show when Arrieta started to rely less on his slider, while the second part will discuss the stats that may explain why he relied less on the pitch.
Arrieta relied less on sliders in 2016
As the 2016 season progressed, Arrieta gradually decreased his slider usage. I was surprised to find that he was getting away from throwing one of the league’s most devastating offerings. Until, that is, I saw that the once effective pitch had started to become less so. The brilliant Jeff Sullivan of FanGraphs beautifully illustrated the deterioration of Arrieta’s slider.
Perhaps the decline in the slider’s effectiveness can be attributed to less movement. For instance, Arrieta typically averaged about 3.5 inches of cutting action, but he was hovering around only 3 inches for the better parts of May, June, July, and August, as illustrated below.
We could probably stop the post here since we’ve established that Arrieta started throwing flatter sliders, which led to hitters making better contact, which led him to throw the pitch less often. But riddle me this: How can a once dominant pitch disappear so rapidly? That is something that I still don’t know, but think I have a decent answer.
Why did Arrieta rely less on the slider?
I tried to put myself into the mind of Jake Arrieta. The first thing I realized was how cool it is to finally be able to grow a manly beard. But I needed to dig deeper into the pitcher’s psyche.
I should admit that the only experience I have pitching is blowing my little league All-Star game and pitching horribly in MLB The Show. So you may want to take what I say with a grain of salt. However, I do know that in either of the above examples I would throw a pitch less if I wasn’t comfortable mechanically. But how do we measure that?
I can’t think of a better way to operationalize mechanics than to extract the in-game release point data.
In doing so, I found that Arrieta’s slider release point was 29% more inconsistent in 2016 than it had been the previous season. And how do we know that’s valid? The chance of this being a fluke is less than 0.1%, or so my numbers tell me. Let that sink in a bit.
Next, I compared his inconsistent release point to slider movement. It was clear that the more inconsistent Arrieta’s mechanics were, the less his slider moved (see image below; R= -0.47 for you fellow stats nerds). For the rest who skipped class like me to try to get better at MLB The Show, that represents a linear relationship, or the extent to which two variables are related. A -0.47 is mild in general, but bears a little more weight in the baseball world. Again, how do we know that’s valid? There’s less than a 0.5% chance that this is random.
Let’s sum this up
We know for a fact that Arrieta threw fewer sliders — 18.0% in 2016 as compared to 28.9% in 2015 — and we can now see that it’s very possible he was uncomfortable throwing it. We can theorize this by showing that his in-game release point was 29% more inconsistent year-over-year. We also know that hitters were getting better contact on the slider, likely because the pitch was moving roughly half an inch less than his previous average during a large chunk of 2016. And, again, we saw that when Arrieta was more inconsistent with his release point, the pitch moved significantly less.
What is all comes down to is that Arrieta’s slider woes might be attributable to some mechanical flaw that made him more inconsistent with his release point. If he can pinpoint this issue, goodbye, opposing hitters.
*A note on how I arrived at the numbers above (nerd alert)
I downloaded the in-game 95% confidence intervals of Arrieta’s vertical release-point data and horizontal slider movement data from Brooks Baseball. Only 95% confidence intervals were included to ensure the effect was indeed an accurate reflection performance. Data was normally distributed. I computed the release-point interval range and ran a pairwise t-test, comparing his 2015 and 2016 means. There was a significant mean difference in release point variance between 2015 and 2016 (p < .001). I next used a Spearman correlation to examine the association between his slider release point and horizontal movement. The correlation between the two variables was -0.467 ( p< .005).