The Most Improved Pitchers Thus Far by Projected WAR

What follows represents an attempt by the author to utilize the projections available at the site to identify the five starting pitchers whose per-inning WAR projections have most improved since the beginning of the season.

For every pitcher, what I’ve done is first to calculate his preseason (PRE) WAR projection prorated to 150 innings, averaging together Steamer and ZiPS forecasts where both are available. What I’ve done next is to calculate the prorated WAR for every pitcher’s rest-of-season (ROS) WAR projection (again, using both Steamer and ZiPS when available). I’ve then found the difference in prorated WAR between the preseason and rest-of-season projection.

When I attempted a similar exercise last month, I used updated end-of-season projections instead of prorated rest-of-season ones. The advantage of the latter (and why I’m using it here) is that it provides the closest available thing to an estimate of any given player’s current true-talent level — which, reason dictates, is what one requires to best identify those players who have most improved.

Only those pitchers have been considered who (a) are currently on a major-league roster and (b) have recorded at least 20 innings at the major-league level and (c) are expected to work predominantly as a starter for the duration of the season. Note that Projection denotes a composite Steamer and ZiPS projection. PRE denotes the player’s preseason projection; ROS, the rest-of-season projection. Inning estimates for both PRE and ROS projections are taken from relevant pitcher’s depth-chart innings projection. Data is current as of some time in the middle of the night between Monday and Tuesday.

5. Drew Hutchison, RHP, Toronto (Profile)
Projection (PRE): 115 IP, 7.5 K/9, 3.4 BB/9, 1.2 HR/9, 4.38 FIP, 1.1 WAR
Projection (ROS): 132 IP, 8.3 K/9, 3.2 BB/9, 1.1 HR/9, 3.93 FIP, 1.7 WAR

Notes
Hutchinson’s most recent start, a complete-game shutout versus Texas last Friday, is the best he’s produced this season in terms of run prevention (box). That said, the single-game 3.69 xFIP he recorded against the Rangers actually represents the exact median figure among the nine starts he’s made in 2014. Strikeout rate, walk rate, and home-run rate: those are the metrics which most readily correlate with ERA, and Hutchison’s projections have improved by all three measures since the beginning of the season.

4. Aaron Harang, RHP, Atlanta (Profile)
Projection (PRE): 10 IP, 6.5 K/9, 3.2 BB/9, 1.4 HR/9, 4.75 FIP, 0.0 WAR
Projection (ROS): 70 IP, 7.8 K/9, 3.0 BB/9, 1.0 HR/9, 3.92 FIP, 0.5 WAR

Notes
Because his preseason projections were issued while he was still technically a member of the Cleveland Americans and because he’s presently a member of the Atlanta Nationals, Harang’s raw rate stats appear to have improved more than is actually the case. Jared Cross, who is the boss of the Steamer projection system, estimates that a switch from the one league to the other would decrease a pitcher’s ERA forecast by about a 0.5 runs. That said, WAR adjusts both for league and park — and, according to WAR, Harang is likely to produce about 0.65 more wins per every 150 innings now than one would have reasonably expected before the season began.

3. Jesse Chavez, RHP, Oakland (Profile)
Projection (PRE): 131 IP, 7.0 K/9, 2.9 BB/9, 1.1 HR/9, 4.14 FIP, 0.3 WAR
Projection (ROS): 119 IP, 7.7 K/9, 2.8 BB/9, 1.0 HR/9, 3.93 FIP, 0.9 WAR

Notes
As noted with regard to Aaron Harang above, one variable that can influence a pitcher’s raw stats is the league in which he pitches. Another one — one germane to the question of Jesse Chavez and his value — is the role in which he’s utilized by his club. In most cases, moving a pitcher from a starting to a relief role will allow that pitcher to produce better rate stats — because he’s able to throw harder, avoid opposite-handed batters, etc. Conversely, a pitcher moving from the bullpen to the rotation will likely see his rate stats decline. What’s notable about Chavez is that, despite having been projected this preseason to throw a number of innings in relief, he’s managed simultaneously to earn improved projections for his rate stats while also receiving a higher percentage of his projected innings as a starter. His projection has improved by almost a win per every 150 innings.

2. Roenis Elias, LHP, Seattle (Profile)
Projection (PRE): 1 IP, 6.0 K/9, 3.8 BB/9, 1.2 HR/9, 4.82 FIP, 0.0 WAR
Projection (ROS): 77 IP, 7.2 K/9, 3.7 BB/9, 0.9 HR/9, 4.16 FIP, 0.7 WAR

Notes
Entering the season, the left-handed Elias was included neither among Marc Hulet’s organizational top-15 prospect list for Seattle nor among Baseball America’s top-30 list for that same club. Nor had he necessarily given any indication that he was a Tommy Milone sort — which is to say, one who produces superlative minor-league numbers and receives excellent projections while nevertheless failing to impress scouts. That Steamer didn’t forecast him to produce numbers worse than replacement level is the most optimistic comment one could reasonably make. Still, on a club that featured the prospect triumvirate of Danny Hultzen, James Paxton, and Taijuan Walker, it’s been Elias who’s produced the best numbers — with peripherals (20.6% K, 9.4% BB, 51.0% GB) to suggest that the results are sustainable.

1. Dallas Keuchel, LHP, Houston (Profile)
Projection (PRE): 68 IP, 6.0 K/9, 3.0 BB/9, 1.0 HR/9, 4.23 FIP, 0.4 WAR
Projection (ROS): 104 IP, 6.5 K/9, 2.8 BB/9, 0.9 HR/9, 3.93 FIP, 1.1 WAR

Notes
Mike Petriello announced earlier today that Dallas Keuchel is the sort of person who can no longer be ignored. The improvement in Keuchel’s prorated WAR projections would appear to support that sentiment. Through nine starts, the 26-year-old left-hander has already surpassed his WAR total from 2013 by over half a win — in about 90 fewer innings. Even despite the fact that he made approximately a third of his appearances last season in relief (in which role he’d be expected to produce better raw figures), Keuchel has improved his strikeout rate and walk rate, in addition to having induced grounders at a frequency unmatched by literally every other qualified starter in the majors. Because, like Jesse Chavez above, Keuchel was projected to record some of his innings out of the bullpen, his raw rate stats don’t necessarily reflect the full measure of his improvement. He’s really good, is the point.

Three additional notes:

• Sixth among pitchers by the measure used here is Garrett Richards (+0.5 WAR per 150 IP). Calculating any of the pitchers after him would require a sort of due diligence, however, which the author is unwilling to perform.
• Among the most improved pitchers excluded from the top five because they’ve been demoted or otherwise have failed to pass the 20-inning threshold, here are the top three: Arizona’s Michael Bolsinger (+2.1 WAR per 150 IP), the New York Yankees’ Chase Whitley (+1.5), and also Arizona’s Chase Anderson (+0.7).
• White Sox right-hander Felipe Paulino (-1.4 WAR per 150 IP) is the pitcher, among those who’ve recorded at least an inning at the major-league level this year, to have exhibited the greatest decline in prorated projected WAR.