Searching for context with Steven Wright

Few baseball things are more fascinating than a dominant knuckleballer. Obviously, the knuckleball itself is a thing of amusement and beauty...

This Steven Wright knuckleball is unhittable https://t.co/ZEfgEs9f5o

— Sox Lunch (@Soxlunch) May 30, 2016

...but knuckleballers, as a group, do not tend toward sustained greatness. Almost by definition, the knuckleball is adopted by only those pitchers who otherwise don't have much of a shot at the majors; that, combined with the inherent unpredictability of the pitch itself, makes a successful user of the pitch rare and exciting, even beyond the fun of watching them pitch. Or at least I think so.

As a result, 2016 has been a good year, thanks to Red Sox pitcher Steven Wright. Through 13 starts and 89.1 innings (an average of 6.9 innings per start), he's sitting on a 2.22 ERA, the seventh-lowest among qualified starters. That is quite obviously good, but we've been taught to place less emphasis on the runs a pitcher allows than the process they undergo while doing so, thus paying more attention to their abilities than their outcomes. Knuckleballers confound that a bit, because unlike most pitchers, they seemingly do exercise some amount of control over the quality of contact they allow, and even when they're performing well, they strike out fewer batters and walk more than their elite peers. While Wright's FIP and DRA are both still good, at 3.37 and 3.43 respectively, they probably underrate him as a result.

That makes putting his performance into context tough. We can look to other knuckleballers, like R.A. Dickey, who won the Cy Young in 2012 on the back of 233.2 innings with a 2.73 ERA/3.27 FIP/3.52 DRA. Beyond both throwing a pitch they call a knuckleball, however, Dickey and Wright don't have a ton in common. Dickey's 2012 had more strikeouts and fewer walks, and his knuckleball routinely broke 80 mph, while Wright is doing a better job of suppressing home runs and hits on balls in play, and throwing a slower knuckleball, ranging from high-60s to mid-70s. Plus, comparing Wright to one other knuckleballer doesn't exactly give us a lot of context; now, instead of one pitcher without comparisons, we have two, which isn't a huge improvement.

With that in mind, I wanted to develop some comps for Wright. Not ones based on his underlying abilities (as we've discussed, he basically doesn't have any) but on his results, hopefully to provide a handhold when thinking about his performance. I took Wright's performance in four areas — strikeout rate, walk rate, BABIP, and home run rate — and scaled them to league average. Wright's walked more batters than average, and struck out fewer than average, but he's also done a pretty good job of suppressing hits and an unbelievably good job of suppressing home runs.

These are the axes the comps will be generated along — not because they're predictive, necessarily, but because they describe most of the experience of watching a pitcher. If I tell you how someone does in these four categories compared to league average, I feel like you get a sense for who they are and how it feels when they take the mound.

What I did next was take every starting pitcher season with at least 150 innings from 2006 to 2015, and scaled each of those to league average (for their season), then added up the differences between their figure in each category and Wright's figures to create a "similarity score," where lower numbers indicate a better fit. Wright is unique, and none of his comps fit particularly well; Dickey's 2012, for example, has 21 seasons more similar to it than the most similar season to Wright's 2016. Nonetheless, this conveys some information, so here are the top ten:

I should also mention, these numbers aren't adjusted for league or park, so don't go reading any rigor into these comps. (Pause for a moment to appreciate Jason Jennings, who held batters to a home run rate 34% below average over 212 innings in 2006, 101 of which came in Coors Field.) Still, it's an interesting list to be sure. As is expected for pitchers who posted below-average rates of hits and home runs allowed, their RA9-WAR's generally exceed both their fWARs and WARPs. By each measure, however, this is a good group of pitchers, with averages of 4.4, 3.6, and 2.9 respectively. Especially if you give them credit for their contact management, this is a very valuable profile.

That said, it's a group that, even in these seasons when they were performing well, left observers with a vague sense of unease. Maybe that's an unfair use of hindsight, but I think even at the time these pitchers seemed to be walking the knife's edge, and on the verge of collapse at nearly any moment. In that way they're not dissimilar from a knuckleballer. Clay Buchholz has had his ups and downs since his very good 2010, and was recently moved to the bullpen in Boston; Shelby Miller has struggled mightily to open 2016; Matt Cain would hold the magic for just one more year after 2011 before losing it, and he hasn't regained it yet. It's not a list full of one-hit wonders necessarily — Johnny Cueto has a deserved reputation as a FIP-beater, and has consistently been very good — but neither is it a list with a lot of staying power.

I think the real lesson from this list, however, is that knuckleballers really have no comparisons. At this point in time, we've all internalized the lessons of Voros McCracken and DIPS theory, but knuckleballers just don't follow the same rules as everyone else. Of the 967 pitchers in this sample, only five had lower home run rates than Wright, and none of them had a lower BABIP. Normally, this kind of thing feels terrifying, entirely luck-driven and entirely temporary. With a knuckleballer, it's still terrifying, but there's at least a slim chance that it is real, based on skill and permanent enough to last.

Wright's only halfway through a season, as well, so this isn't exactly apples-to-apples. Even with a knuckleball, he's been performing so well that one should probably regress him some amount back towards normal. If we indulge in some linear projecting, however, and assume he finishes the season with a 2.22 ERA, it would be the 10th-best ERA since 2006, behind four Cy Young seasons. Thanks to Danny Salazar giving up a pair of runs on Saturday, Wright also has a slim lead in the AL ERA race. He's pitching in a way that basically nobody pitches, suppressing home runs and hits like mad. If he keeps doing it, Wright will have a strong claim to some accolades.

. . .

Henry Druschel is a Contributing Editor at Beyond the Box Score. You can follow him on Twitter at @henrydruschel.