Ignore the modest save total of twelve: Andrew Miller was one of the best relievers in baseball during the 2016 regular season, and was the best reliever in baseball during the postseason. He was unquestionably the deciding factor in the Blue Jays ALCS defeat at the hands of Cleveland, as favorable scheduling and deft usage by Terry Francona allowed the left hander to appear in four of the five games. During that stretch, he he struck out 14 of his MLB-record 30 postseason batters over 7.2 innings of work.
Mark Shapiro, Ross Atkins, and the rest of the front office would certainly love to get their hands on Miller, but with two reasonably priced years remaining on his contract Cleveland seems unlikely to bring him to market – and even if they did, it would be for an unpalatable ransom for a reliever coming off a massive single-season workload of 93.2 innings. While in possession of similarly dominant track records, even if one could stomach the estimated 15 million (or more) per season they’ll demand, Kenley Jansen and Aroldis Chapman seem like equally far-fetched alternatives due to their attached qualifying offer and history of domestic violence, respectively. With all due respect to Brett Cecil, who is a very, very good reliever in his own right, if the Blue Jays want to add a dominant reliever to the 2017 bullpen without crippling their payroll, farm system, or morals, they’re going to need to get creative.
So creative, in fact, they might want to look backwards before looking forward.
Long before Andrew Miller took over the baseball world with a seemingly infinite cascade of wipe-out sliders, he was a starting pitcher for the Detroit Tigers, Florida Marlins, and Boston Red Sox. Miller was taken sixth overall in the 2006 amateur draft, debuted that same year, and after dabbling with a cutter and curveball in the early stages of his career, settled into a three pitch mix featuring his four seam fastball, slider, and changeup.
While he had the occasional bullpen stint, Miller was used primarily as a starting pitcher over the next five seasons but never performed particularly well despite the flashy pedigree and substantial signing bonus. The left hander would be traded twice; the first time headlining the package for Miguel Cabrera, the second time for someone named Dustin Richardson in a straight one-for-one before being non-tendered shortly thereafter, perfectly epitomizing how far his stock had fallen by the end of 2010.
Red Sox manager Terry Francona pushed to have the club bring Miller back after he was non-tendered, as he and the coaching staff felt as though they could correct a flaw in his mechanics and get his career back on track. A month later Miller came to terms with the club, but despite Francona’s aspirations, the 2011 season was yet another tumultuous adventure through the wilderness for the lefty. In one of his final moves before Boston hilariously moved on to Bobby Valentine at the end of the 2011 season, Tito made the call to transition Miller to the bullpen full time – a decision that seems absolutely brilliant in hindsight.
The first thing that Andrew Miller did as a reliever was ditch his changeup, a dreadful pitch that had done nothing to stifle right handed batters throughout his career. He had thrown the changeup around 15 percent of the time over three years from 2009 through 2011 – more often than his slider in the 2009 and 2010 seasons, in fact. In its stead, Miller immediately began throwing that slider 40 percent of the time, and has since upped that usage as high as 60 percent in the 2016 season. Conversely, Miller’s slider is obviously a fantastic pitch. Even at his lowest point the breaking ball was getting whiffs on over 35 percent of swings taken against it, and more recently has generated whiff per swing rates of 55.1, 55.2, and 46.9 percent over the last three years.
Miller’s four seam fastball received the “bullpen bump” after he transitioned to relief. The pitch averaged between 91 and 92 miles per hour in the rotation in 2009 and 2010, climbed over 93 in 2011 thanks to a big September, and has averaged between 95 and 96 miles per hour in each of his five full-time relief seasons. A lot is made of velocity – admittedly sometimes too much – but as Table 1 explains, the move allowed Miller to cross a critical velocity threshold.
The graph above was created using the PitchFX data available on MLB Savant, and shows how batters perform against four seam, two seam, and sinking fastballs as pitch velocity increases. Both batting average and slugging percentage decline as velocity increases, which is expected, but what is most interesting is how the velocities appear to coalesce into tiers. Fastballs from 88 to 91 miles per hour are damaged at more-or-less the same magnitude, but there’s a sharp decline in batter performance from there to 93 miles per hour, where batting average and slugging percentage drop by about 30 and 70 points, respectively. The 93 to 95 miles per hour window is again something of a plateau, with a steep deterioration in batter success at every step from 96 and beyond. In moving from the rotation to the bullpen, Miller’s four seam fastball jumped more than a full tier, with the expected slugging percentage falling from .495 to .420. That’s monumental.
With this information in hand I set out with the goal of finding the next Andrew Miller, or at least, a poor man’s Andrew Miller. Using the pitching leaderboard over at Fangraphs, I ran a report for all pitchers who threw 30 or more innings in the rotation in 2016, which turned out 213 names. From there, I began trimming the sample with criteria that meet or best describe Andrew Miller’s transformation. Pitchers who were worth more than 1.5 fWAR in 2016 were removed first, as the foundation for this exercise is to look at failed starters. From there, the population was cut further to only those pitchers with an average fastball velocity of 91.0 miles per hour or greater in order for the “bullpen bump” to facilitate a leap into the next velocity tier. The third and most pivotal step was to eliminate all pitchers that did not have two or more secondary pitches thrown in excess of 10 percent of the time, as like Miller and his changeup, there must be something to dispose of at the expense of better offerings.
This trimmed the list of pitchers from 213 down to 47, which was reduced further by removing the “obviously not going to be converted” starters that also met the criteria such as Sonny Gray, Jose Berrios, Daniel Norris, and Garrett Richards among many others. Now down to 30, a few names started to jump out as intriguing prospects for a bullpen conversion. Jimmy Nelson, Clayton Richard, and Lucas Harrell all have a decent fastball/breaking ball combination and a changeup just begging to be dropped, but one name truly emerged as an almost perfect candidate to be the next Andrew Miller: Patrick Corbin of the Arizona Diamondbacks.
The left hander made 24 starts from April through mid-August, over which he struggled to a 5.58 ERA and equally troubling fielding independent metrics. After giving up eight runs on nine hits in what would be his final start of the year on August 12th, the Diamondbacks made the decision to shift him to the bullpen for the remainder of the season. With a 2.70 ERA over 12 appearances his performance was nothing short of spectacular, but even that fails to give proper justice to how dominating Corbin was down the stretch, and how well his arsenal played up in the bullpen.
Despite it being the tail end of his first full season back from Tommy John surgery, the “bullpen bump” was in full effect for Corbin. His four seam fastball and sinker, which averaged 92.3 and 93.6 miles per hour in the rotation, jumped all the way up to 94.2 and 94.6 miles per hour in the bullpen; a surge of nearly two full ticks on his primary offering. Referring to Table 1 above once again, while this increase doesn’t appear to do a ton to stifle opposing batting averages, jumping from 92 to 94 carries an expected decline in opposing slugging of 37 points. The sample size is admittedly small, but this played out on the field as batters slugged .565 against his four seam fastball as a starter and just .452 as a reliever.
In relief, while his fastball (four seam + sinker) usage remained steady at roughly 63 percent, Corbin dropped his changeup usage from 10.46 to 5.11 percent. This decline was replaced by a growth in his slider reliance, which increased from 25.65 to 31.82 percent of total pitches thrown. That should be taken a whole lot further, too.
Andrew Miller’s slider is lauded for being nigh-untouchable, and rightfully so. What’s truly remarkable about Patrick Corbin is that his slider might be just as good, if not even a little better than that of Miller. Table 2 below details Miller and Corbin’s slider usage and whiff rates over the last five seasons. Despite facing batters two, three, and even four times on any given night as a starter, Corbin’s slider produced a better whiff per swing rate than Miller in two of his three healthy seasons. When the playing field was finally leveled as Corbin shifted to the bullpen at the end of 2016, his whiff rate blew past Miller and it wasn’t all that close.
That there is one bastard of a breaking ball. Using the Baseball Prospectus PitchFX Leaderboard, its whiff per swing rate of 61.36 percent was the second highest among 165 relievers to throw 100 or more sliders, behind only Ken Giles of the Astros, and its groundball rate of 70 percent was tied for first. Patrick Corbin’s slider is legitimately one of the best pitches in all of baseball, and it has been criminally underutilized for far too long.
Honestly, it’s legitimately dumbfounding that Corbin has continued to throw a changeup at all. Its primary utility is to shut down opposite handed batters, especially for pitchers who feature a slider associated with heavy platoon splits. It hasn’t played out that way for Corbin. The lefty has thrown 886 changeups to right handed batters in his career, and they’ve demolished it for a .335 batting average and .608 slugging percentage. Meanwhile, the 1,326 sliders that Corbin has thrown to right handed batters have resulted in a .145 average and .256 slugging percentage.
Again, versus right handers, he’s surrendered just eight home runs with 224 strikeouts with the slider, compared to eleven dingers and just 10 strikeouts with the changeup. Moving forward, regardless of his role — though it would obviously play up better in the bullpen — Patrick Corbin should be a four seam fastball, sinker, and slider guy. Full stop.
If the elite slider alone isn’t enough to whet your appetite, what makes Corbin even more fascinating as a relief option is the length that his history in the rotation could allow. It was noted earlier that Corbin made 12 relief appearances to conclude the season, but what was intentionally omitted at the time was that those outings covered 23.1 innings. Excluding the August 19th appearance in which he was shelled and the September 9th appearance in which he was asked to face only one batter, the following are the number of batters Corbin faced in each of his other ten outings: 8, 9, 10, 6, 8, 12, 9, 8, 7, and 9. In a surprisingly progressive fashion given the organization’s previous hierarchy, Diamondbacks manager Chip Hale basically asked Corbin to blow through the entire lineup one time, or until his own spot came up in the batting order.
If you forecast his workload from the final seven weeks out to a full season, Corbin’s relief usage would look something like 87 innings spanning 45 appearances. Brad Hand of the Padres led all of baseball with 89.1 innings of work out of the bullpen in 2016, so while using Corbin in that way wouldn’t be unprecedented; it would certainly fall near the current extremes. The biggest difference between the two would be that it took Hand a rather questionable and potentially abusive 82 appearances to get there.
Having a stable of elite relief options was in vogue during the Royals 2014 and 2015 runs to the World Series, but the 2016 postseason took things a step further as managers turned to those elite options earlier and more often. It feels like we’ve experienced a shift in philosophy; away from catering to the arbitrary “save” statistic and towards the far more logical “use your best damn guys in the biggest situations” approach.
It may seem like he’s been around for a while but Corbin will still be just 27 years old on Opening Day, has two years of team control remaining, and should make in the neighborhood of 4.2 million next year according to the MLB Trade Rumors arbitration estimates. Rather than trying to reinforce the bullpen by spending top dollar or top prospects, the Blue Jays might be better served to follow this or a similar exercise, look at how the best arms came to be the best arms, and work from there to recreate them. It seems obvious, but if you want an Andrew Miller but don’t want to pay Andrew Miller prices, find the next Andrew Miller. Patrick Corbin might just be that guy.
*All PitchFX figures from BrooksBaseball.net unless otherwise noted.
Lead Photo: Matt Kartozian-USA TODAY Sports