Hoskins, Castillo, and Quantity vs Quality in Awards Voting

Last night, the Rookie of the Year awards were announced, with Cody Bellinger and Aaron Judge getting every first place vote in their respective league, as expected. The rest of the ballots were more interesting, with plenty of options for second and third place in both leagues. Eno posted his NL ROY ballot last night, explaining why he went with Rhys Hoskins and Paul Dejong as his post-Bellinger votes.

I also had an NL ROY ballot this year, but it differed from what Eno turned in, and in fact, differed from what everyone else turned in too. I was the only voter to include Reds RHP Luis Castillo on a ballot, as I put him third behind Bellinger and Dejong, leaving off Hoskins, among others. And while I know down-ballot Rookie of the Year voting isn’t the most exciting thing going on right now, I think it is useful to use that vote as a way to think about how we balance quantity and quality when determining past value.

First, I should note that I’m not convinced my ballot was “right”, in the sense that there can be correct votes. I considered a bunch of guys for the last few spots on the ballot, with Hoskins, Manuel Margot, and German Marquez getting the most consideration of the guys who just missed the cut. A lot of other people voted for one or two of those three — and Josh Bell, who I didn’t seriously consider given that he wasn’t actually very good in 2017 — and the non-Bellinger guys mostly have similar enough cases that a vote for any of them isn’t any kind of egregious mistake.

But I will tell you why I ended up with Dejong and Castillo behind Bellinger, despite Hoskins’ excellence after his promotion and my general affinity for his future.

Dejong probably requires a little less explanation, as he appeared on 26 of the 30 ballots, and pretty easily finished second, so my placement there was in line with the consensus. Dejong didn’t hit as well as Hoskins, of course, but he spent four months in the Majors instead of just two, and critically, he played shortstop instead of faking it in left field. The only shortstops who ran a higher wRC+ and came to bat at least 400 times were Carlos Correa, Zack Cozart, and Corey Seager. Dejong was really good, and played most of the season, so putting him second on my ballot wasn’t that tough a choice.

In the end, it really came down to Castillo and Hoskins, both of whom played partial seasons and were excellent in the time they were given. Among batters with at least 200 PA, Hoskins’ .417 wOBA ranked fifth in MLB. Among starting pitchers with at least 80 innings pitched, Castillo’s .276 wOBA allowed ranked 10th in MLB. And Castillo’s xwOBA, which includes batted ball quality, was actually a ridiculous .257, ranking fifth among MLB starting pitchers.

Both were elite players for a chunk of the season, and both are deserving of recognition.

So why did I end up with Castillo? While Eno focused on “total chances”, Castillo started about half the games of a normal starting pitcher, while Hoskins only started about one-third the games of an everyday position player. It’s true that position players are generally more valuable than pitchers, given that they contribute both on offense and defense, but given that Hoskins was a first baseman playing left field, I don’t think he added a tremendous about of value on the fielding side of things, leaving us to mostly focus on his offensive value versus Castillo’s pitching value.

Eno noted that Hoskins wRC+ was 58 points better than average, while Castillo’s ERA- was only 29 points better than average. But while both metrics are centered around 100, I think we need to keep in mind that a point of wRC+ isn’t equal in value to a point of ERA- (or even FIP-/xFIP-), and the spread in talent at those skills isn’t equally distributed. For Castillo to match Hoskins’ 158 wRC+ just in distance from 100, he’d have to have put up a 42 ERA-, which has been matched by exactly three pitcher seasons (out of 1,498 attempts) with 80+ innings since 2008. A 158 wRC+ or better in a season with 200+ PAs has been accomplished 65 times in 3,497 player seasons over the last decade.

Setting a point distance from average underestimates how much more difficult it is to prevent runs over a half-season than it is to create them over a third of the season. So, in the end, I decided that Castillo’s quantity advantage slightly outweighed Hoskins’ production advantage, as both were similarly elite producers in their time in the Majors, but with Castillo doing it over a larger sample.

At the end of the day, third place votes on Rookie of the Year ballots are pretty inconsequential, but we’ll probably have similar quality-versus-quantity discussions when it comes to the Cy Young and MVP awards, and working through the ways we weigh elite performances in different sized samples can help in those discussions as well. And if Castillo and Hoskins build off what they did in their rookie seasons, this won’t be the last time we’re talking about either one when it comes to awards season.