Felix Hernandez and Situational Pitching

We’ve got a lot of weeks ahead of debating Felix Hernandez vs. Corey Kluber, as the Cy Young race seems like it ought to be a dead heat. Several different elements come into play, and to be perfectly honest it’s basically impossible to separate the two from one another, but something that’ll get talked about is Felix’s ERA advantage. While Kluber has outstanding peripherals, people also care about actual runs, and Kluber’s allowed a dozen more runs than Felix has. Some of this is probably because of defense. Some of this is probably because of ballpark environment. But you also have to consider this:

Felix Hernandez 2014 splits

Bases empty: .259 wOBA allowed
Runner(s) on: .215
Scoring position: .208

In run-scoring opportunities, Felix this year has stepped up his game. With the bases empty, that wOBA allowed ranks tied for 22nd. With runners on, that wOBA allowed ranks first, and by a whole 17 points. As you could guess, this sort of thing needs to be regressed, and it won’t surprise you to learn that Felix’s BABIP is also lower in run-scoring situations. But this goes beyond just a BABIP thing. Felix probably deserves some credit for this, because that doesn’t all seem like a fluke.

On the year, 154 pitchers have thrown at least 40 innings with the bases empty and with men on. With men on, they average a worse strikeout rate by 1.4 percentage points, and a worse walk rate by 1.1 percentage points. So they average a worse FIP by 0.22 points. How about Felix? His strikeout rate has been better by 7.4 percentage points. His walk rate has been better by 1.0 percentage points. His FIP has been better by 0.97 points. So it’s not just the wOBA that’s improved — it’s supported by more sustainable statistics. One of the reasons Felix has generated better results with men on is because he’s turned in better pitching. It’s a pretty simple explanation, that one.

So Felix has put up better fielding-independent numbers when runs are more likely to score. That’s interesting — that indicates a different process — but can we find anything within the actual processes? As a matter of fact, we can, and it doesn’t even take too much work. All it takes is some time spent at Brooks Baseball.

What does Felix throw? He throws a bunch of pitches, but people know him best for his changeup. He might have the best changeup in baseball, and he’s been in love with it ever since 2009. He’s not the least bit afraid to throw it to righties, and it’s one of baseball’s true dominant pitches, the kind of pitch a guy can’t hit even if he knows that it’s coming. Here is a plot of situational changeup rates, by the Brooks numbers.

Every season, Felix has thrown a higher rate of changeups with men on base. But the gap has widened. Year by year:

• 2009: 3% difference (percentage points)
• 2010: 5%
• 2011: 7%
• 2012: 6%
• 2013: 9%
• 2014: 16%

This season, with nobody on, Felix has thrown a changeup about a quarter of the time. If there’s been someone on base, though, he’s thrown a changeup more than two-fifths of the time. I understand that Felix’s changeup and fastballs can be difficult to separate in the data, so perhaps this comes with bigger error bars than usual, but I trust the Brooks system, and these numbers help to explain the other numbers observable within Felix’s splits.

See, it stands to reason, when a guy throws more of his most unhittable pitch, he’ll get more unhittable results. The question might be, then, why not pitch like this all the time? And I don’t know the answer to that, but maybe it’s a little extra stressful to throw more changeups. Maybe Felix doesn’t want hitters to see too many of those pitches until they matter most. The fact that the numbers are what they are over a big sample suggests that this is a conscious thing, and while I don’t know the correct answer as to why this is, the explanation is less important than the fact. Maybe it’s just easier to get hitters to over-commit when they’re thinking about driving someone in.

Here’s what happens to Felix’s location, courtesy of Baseball Savant. One image shows his pitches with the bases empty; the other shows his pitches with the bases less empty. The latter is the image with the pitches clustered a little bit lower, since the changeup is a pitch meant to be thrown around the lower edge or below.

Felix’s average pitch with runners on is almost two inches lower than his average pitch with the bases empty. With no one on, Felix has thrown 71% of his pitches no more than 2.5 feet off the ground. With someone on, that goes up to 78%. That’s not the most dramatic difference in the world, but it’s something, and again, this isn’t just about location — it’s also about timing. It’s about whatever it is that makes Felix’s changeup so consistently lethal. When batters have been trying to bring runners home, Felix has countered that by throwing his elite pitch 42% of the time, and the results to this point have rewarded Felix’s situational thinking.

The actual effect of all this on Felix’s ERA is less significant than you might think. We’re really just talking about a handful of runs, and Felix is amazing regardless of his approach. If he does win the Cy Young, it won’t just be because he pitched better in more nervous situations. But that’s going to be a factor, directly or indirectly, and it seems like it’s not just all a fluke. It seems like Felix has intended this, to some extent, and so he’s turned in some more timely strikeouts, and he’s avoided some more untimely hits. Absolutely, Corey Kluber has a case for the AL Cy Young. Absolutely, there are reasons Felix remains the favorite.