Jays From the Couch takes a deep dive into the Blue Jays’ rotation and their ability to manage contact
I should get this out of the way immediately: none of the projection systems say that the Jays have a Top 10 rotation (to the best of my knowledge). And that’s okay. When thinking about projections, it’s necessary to talk about the assumptions that drive them (like Stoeten did with the team’s PECOTA projections). That’s something I intend to do in more detail with the Fangraphs projections once the Opening Day roster is finalized. Instead, today I’m going to offer a take on why the Jays’ rotation should be considered amongst those just behind the very best in MLB.
The Blue Jays have developed a reputation as a team with a pitching staff stacked with excellent contact managers. In a recent (must-read) post, Joshua Howsam broke down exactly how each pitcher accomplishes their particular brand of weak contact, from the ground-ballers (Jaime Garcia, Marcus Stroman and Aaron Sanchez) to the weak fly-ballers (Marco Estrada). Unfortunately for analytic-minded Jays fans, staffs loaded with elite contact managers are often projected inaccurately, as projection systems still struggle with pitchers who generate weak contact.
Those pitchers tend to have an unusually low BABIP (since a disproportionate number of balls in play turn into outs) and an ERA lower than their FIP (since they aren’t usually also strong at generating strikeouts and limiting walks). They are “punished” for it by systems that assume a pitcher will have 1) a roughly league-average BABIP in the long-run and 2) an ERA and FIP that are equal to each other.
So, in spite of the fact that the Jays’ 2018 starting pitchers have cumulatively maintained an average BABIP of .287 over the last three seasons (weighted by their projected innings in 2018), Fangraphs projects them to produce a mark of .306 in 2018. And, even though their average ERA has ran a bit lower than their average FIP, Fangraphs assumes the two measures will be the same in 2018. The result is a hypothetical jump in the rotation’s ERA of more than half a run, two fewer fWAR and six fewer wins above replacement in terms of actual runs allowed.
The approach I’ll discuss in this post is one (imperfect) way to compare rotations, whether they excel at contact management or generating strikeouts and limiting walks. This approach estimates that the Jays’ 2018 starting rotation ranks tenth in the majors (that still counts as Top 10, so my click-bait title is technically accurate).
Before we get into the results, let me explain my methodology a little bit more. Fangraphs’ pitcher WAR projections are based on FIP, which focuses on a pitcher’s ability to generate strikeouts, while limiting homers and walks. Their ability to avoid allowing other types of base-hits is left out of the equation, as pitchers generally have limited control over the outcome of a ball in play. Generally, this is fine, best shown by the fact that a pitcher’s past FIP is more predictive of his future ERA than his past ERA is. However, one of the key limitations of FIP is that some pitchers do have a proven ability—an ability to generate weak contact more often than most pitchers—to limit the likelihood that a ball in play will become a hit.
There is a ready-made solution to this issue, which will surely be a part of a public projection system once enough data has been collected—xwOBA. xwOBA (public data goes back to 2015) accounts for a pitcher’s actual number of strikeouts and walks, but replaces actual batted ball outcomes (singles, doubles, triple, homers and outs) with projected batted ball outcomes based on the exit velocity-launch angle combination for each ball in play. For example, a ball hit with an exit velocity of 100 mph and a launch angle of 30 degrees turned into a single 0% of the time, a double 9% of the time, a triple 6% of the time and a homer 47% of the time. It would have an xwOBA of 1.199.
On the other hand, a ball hit with an exit velocity of only 70 mph and a launch angle of -20 degrees turned into a single 4% of the time (but is highly unlikely to become an extra base-hit). It would have an xwOBA of 0.053. [I have no idea why Statcast uses Progressive Field as their default field for this tool.]
This is a metric that gives pitchers credit for the kind of contact they give up. So, like FIP, xwOBA for pitchers is effectively fielding and base runner independent—it doesn’t punish pitchers for their fielders’ limitations or the speed of opposing batters. But, unlike FIP, it does reward good contact managers, who consistently generate slow exit velocities and extreme launch angles (getting batters to weakly top or get under their pitches).
For the purpose of this post, I estimated an xwOBA for each team’s 2018 starting rotation. In my calculations, I included any starting pitcher who 1) is projected by Fangraphs’ Depth Charts projection system to pitch 50+ innings as a starter in 2018 and 2) pitched 50+ innings as a starter over the 2015-17 seasons (all of the seasons in which xwOBA data is available). These are my “Qualified SP”. Then, I calculated a mean xwOBA for each team using the 2015-17 xwOBA of their qualified starters, weighted by each pitcher’s projected 2018 innings. I repeated this for pitchers’ xwOBA on batted balls only, strikeout rate and walk rate. [I hope I haven’t lost my readers by this point.]
For most stats, a minimum sample size of 50 innings is probably too small. For our exercise, the benefits of doing so—a 50 inning minimum ensures that all teams have at least four qualified SP—far outweigh the costs—strikeout and walk rates are reasonably accurate measures of performance by 50 IP, while xwOBA is inherently useful in small sample sizes.
Also, to be clear, this is not a projection. But, it is a way for us to think about the future in a way you otherwise can’t—with a pitcher’s contact management accounted for.
The table below includes the number of qualified SP per team, their estimated xwOBA on batted balls only, their strikeout and walk rates, and their total xwOBA (which essentially brings together the first three statistics into one metric). As my headline promised, this methodology ranks the Jays tenth in the majors in starting pitching.
In jumping from 14th (in projected WAR) to 10th (in xwOBA), the Jays rotation leapfrogged the Rays, Rockies, Diamondbacks and Pirates. The leap forward was propelled by the much better-than-average xwOBA of Estrada (.296), Garcia (.303) and Joe Biagini (.303), relative to their very limited projected WAR (1.1, 1.4 and 1.2). On the other hand, Stroman’s projected WAR (3.9) is more bullish than his xwOBA (.310).
Chris Archer‘s xwOBA is in the Top 15% (.290), while his projected WAR is in the Top 5% (4.4). As a result, by xwOBA, the Rays’ projected WAR is “overstated” by about 1.5 WAR, enough to put them past the Jays rotation. In the case of the Rockies, it’s Jon Gray whose projected WAR is overstated by one and a half wins, relative to his xwOBA.
The Diamondbacks have three starting pitchers (Robbie Ray, Patrick Corbin and Taijuan Walker) with high WAR marks relative to their xwOBA, for a total gap of nearly 3 WAR. Finally, for the Pirates, five pitchers were overstated by a total of 5 WAR relative to xwOBA: Jameson Taillon, Joe Musgrove, Ivan Nova, Chad Kuhl and Tyler Glasnow. [None of this is to say that these players are genuinely worse than projected. There are reasons other than talent to explain why their projected WAR is relatively stronger than their 2015-17 xwOBA. Instead, it’s intended to explain the movements up and down the rankings.]
A good test of an exercise like this is whether the results are fairly intuitive or completely nonsensical. These results seem fairly intuitive. The Top 9 teams by xwOBA include nine of the Top 10 teams by projected WAR. The Dodgers have a fantastic rotation, the White Sox do not. Overall, a rotation’s past xwOBA is highly correlated with its projected WAR (R2 of 0.74), with an extra .010 of rotation xwOBA associated with a 2.5 WAR drop in its projection.
One issue with using past xwOBA to think about future performance is the fact that rookies don’t have a past xwOBA. That particularly hurts the Angels, given that Shohei Ohtani represents a quarter of their rotation’s projected WAR. However, the team is only penalized for having a good pitcher excluded, as the innings that are excluded aren’t included in the team xwOBA calculations.
It’s clear how the Jays rotation found itself in the ten spot—it is elite at limiting dangerous contact. Over the last three seasons, each of the six starters projected to crack 50 IP in 2018 have generated an xwOBA on batted balls less than the 2015-17 MLB average. On the other hand, only Estrada and Happ cracked the league-average strikeout rate, while only Stroman and Happ cracked the league-average walk rate. That said, the group’s strikeout and walk rates are collectively pretty close to the MLB average for starting pitchers.
Positively, five of the six starters (excluding Sanchez) produced xwOBA over the last three seasons that were comfortably better than the league average. In Sanchez’s case, his 2015-17 mark is weighed down by his rocky rookie season (.343 xwOBA) and injury-blighted 2017 campaign (.361 xwOBA), while he was better-than-average during his breakout season in 2016 (.314). Obviously, Jays fans are hopeful that his true talent level is closer to 2016 than 2015 or 2017. That seems highly likely to me. While it’s fair to say that Sanchez’s elite 2016 results (his 3.00 ERA, for example) were certainly driven, in part, by good luck—he posted the largest xwOBA-wOBA among 160 starters who faced 250+ batters that year—his slightly better-than-average performance in 2016 (captured by his xwOBA) seems very doable for him to replicate in 2018.
All in all, you don’t have to squint too hard to see why the Jays’ starting pitching could be a position of strength for the team in 2018. As a group, they strike out and walk batters at roughly league-average rates, while limiting batters to some of the weakest contact in the majors (surpassed only by the super-rotations of the Dodgers, Astros and Nationals). It might not be as sexy (or as rewarded by projection systems) as swing and miss pitching is, but it does the job.
*Featured Image Courtesy Of DaveMe Images. Prints Available For Purchase.
HEAD ON OVER TO THE JAYS FROM THE COUCH VS ALS STORE AND GET SOME GREAT SWAG THAT YOU WILL LOOK GREAT IN AND YOU CAN FEEL GREAT ABOUT.
YOU CAN ALSO HEAD TO OUR JAYS FROM THE COUCH VS ALS FUNDRAISING PAGE TO MAKE A TAX DEDUCTIBLE DONATION DIRECTLY TO ALS CANADA.
THANK YOU FOR VISITING JAYS FROM THE COUCH! CHECK US OUT ON TWITTER @JAYSFROMCOUCH AND INSTAGRAM. LIKE US FACEBOOK. BE SURE TO CATCH THE LATEST FROM JAYS FROM THE COUCH RADIO AND SUBSCRIBE TO OUR YOUTUBE CHANNEL!