Cy Young Voting and the Impact of a Catcher by Dave Cameron September 24, 2015 I have an NL Cy Young ballot this year, and with 10 days left in the regular season, I legitimately don’t know who I’m going to vote for. If I was the kind of voter who looked only at run prevention and credited that entirely to the pitcher on the mound, it might be a fairly easy decision in favor of Zack Greinke, given that he’s at +9.4 RA9-WAR, putting him in pretty historic company in terms of keeping runs off the board. If I went solely by the measures that we have a pretty good idea are primarily influenced by the pitcher and not his defenders behind him, it would be pretty easy to cast a vote for Clayton Kershaw, given that he’s leading the fielding in FIP-based WAR by a pretty good margin. But as we’ve discussed many times here over the years, both of those extremes are clearly incorrect; giving a pitcher no credit or blame for all the non-HR contact he allows is definitely wrong, but so is assigning the entirety of the results of those balls in play to him and pretending that defense does not play a role in run prevention. In reality, a pitcher should get credit (or blame) for some of the impact of events that FIP does not capture, so FIP-based WAR is definitely an incomplete measure of a pitcher’s performance. How much credit or blame should be assigned isn’t entirely clear, and if you just throw your hands up in the air and split the credit down the middle — blending FIP-based and RA9-based WAR together — you’ll note that Jake Arrieta ends up in the top spot, though the differences between all three pitchers at that point are so small as to be insignificant. As you probably know if you’re reading FanGraphs, I’m not going to simply cast my vote based on total run prevention, since I believe in attempting to isolate player performance when handing out individual awards, and simply using run prevention metrics and pretending like defense isn’t a thing strikes me as a particularly lazy shortcut. But I’m also not going to just use FIP, and not just because it ignores a bunch of plays that do matter; it also has (albeit to a smaller degree) issues with teammate interaction. As our ability to measure a catcher’s impact on balls and strikes has grown, it has become clear that no pitching event is really “fielding independent”, and a pitcher’s walks, strikeouts, and home run rates are indeed impacted by the performance of the guy he’s throwing the ball to. Just as we attempt to adjust for things like park factors, knowing that the dimensions of his home field and the weather he pitched in are variables that impact results but are out of his control, it makes sense to also account for the receiving ability of the guy a pitcher is throwing to. If one pitcher is consistently throwing pitches on the corner and not getting calls because his team’s catcher isn’t that great at fooling umpires, do we really want to hold those pitches against him? Or if he’s getting called strikes on pitches well below the knees because of the skill of his catcher, do we want to give him 100% of the credit for those strikeouts? In theory, catcher framing is another variable to attempt to control for, and in this particular instance, it could indeed end up making a bit of a difference. For instance, here are the batters faced totals, by catcher, for the two Dodger pitchers in the Cy Young race: Catcher Allocation Pitcher Grandal Ellis Grandal% Ellis% Clayton Kershaw 329 497 40% 60% Zack Greinke 571 171 77% 23% While Kershaw has had a more even split of the two Dodgers catchers, with a slight lean towards Ellis, Greinke has primarily been caught by Yasmani Grandal. And by nearly every framing metric out there, Grandal is one of the best catchers in baseball at stealing strikes for his pitcher, so it’s likely that Greinke has benefited to some degree by pitching to a high-quality receiver, while Kershaw has probably not gotten as much support from his catchers. And StatCorner’s data on their pitching pages backs this up to some extent, as Greinke is listed to have 8.6% of his out-of-zone pitches taken for strikes, while Kershaw is only at 7.6%. So, on the surface, this is something we’d want to try and account for, which would boost Kershaw’s numbers relative to Greinke’s. But the more I think about this, the less sure I am that this is something we should actually assume is an outside variable, because the identify of the catcher on the days a specific pitcher takes the mound is not random or outside the pitcher’s control. It is well known that Kershaw loves throwing to Ellis, and that the two are very close friends. After Kershaw reportedly asked for Ellis to catch him in his first start of the year, Ellis said this: “It’s really special that (Kershaw) enjoys throwing to me,” Ellis said. “We have a good rapport, good chemistry and we kind of lean on each other out there. If I can add my two percent to his day of making it great, that’s pretty much all he really needs. It’s special when I can call that one or two pitches an outing that’s a separator for us.” The Dodgers didn’t allow Kershaw to have Ellis as his personal catcher this year, but it’s not a coincidence that 18 of the 49 games Ellis has started this year have come when Kershaw is on the mound. So, unlike a park effect or the performance of the bullpen in games he leaves runners on when removed mid-inning, we can’t really say that this is a variable outside of Kershaw’s control. His preference to throw to Ellis is his choice, and if he gains some extra comfort in throwing to Ellis instead of Grandal, the trade-off he’s making is getting fewer borderline pitches called strikes. But he’s making that choice, so giving him a boost for having to throw to an inferior defensive catcher doesn’t seem right. And, to take this a step farther, Kershaw’s preference for Ellis actually degrades the Dodgers offense on the nights he pitches as well; Grandal has cooled off a bit lately, but he’s still posted a 123 wRC+ this season, a bit better than the 109 wRC+ Ellis is running. It’s not a huge difference, but Kershaw’s catcher preference does force the team to put a line-up on the field that is less likely to score runs; perhaps this should be a small negative against Kershaw as well? Or maybe it should be no factor, since Ellis was going to have to draw some starts regardless, and the distribution of when he starts is mostly irrelevant; in this instance, Kershaw would essentially be sacrificing his own numbers by taking Ellis’ starts, but the overall net effect on the team would be close to zero. In reality, these factors aren’t going to move the needle dramatically, but since I’m not judging by either ERA or FIP, I see this as a very close race between three pitchers all having excellent seasons, so my vote could end going one way or another based on a small factor here or there. I’m doing my best to try and explore every possible avenue for determining who really had the best season, including considering the value a pitcher puts up at the plate on days he pitches. And, at this point, I don’t think we have really clear answers on a lot of these smaller variables; we’re still struggling with the big variable of how to handle non-HR contact. So, I’ll put this question to you guys. Should the identity of the catcher that Kershaw and Greinke have thrown to be a factor, even a small one, in determining my Cy Young vote? Should I penalize Greinke for getting to throw to the better-framing catcher more often, or should I penalize Kershaw for having a preference for pitching to the guy we think is an inferior defender, based on what we think we can measure at this point in time? Or should we just admit that there is so much going on beyond simply framing marginal pitches that we can’t really quantify the impact of a catcher on a pitcher’s results, and I should just go back to focusing on how to distribute credit for Greinke’s .180 BABIP with runners in scoring position? I have 10 days to figure this out. I’m pretty sure it’s not going to be an easy call, as even going down paths like this one tend to lead to muddy conclusions with no obvious right answer, at least to me. I think the only thing I can say with any real certainty is that evaluating pitching is still really hard.