May 29, 2017

Statistically Speaking: Detwiler falls short

The decision to move lefty Ross Detwiler to the bullpen wasn’t one made lightly or with much joy by the Washington Nationals. Beset by injury and a 2014 spring training showing that found him working through a rough patch with respect to harnessing his secondary pitches (including a new one, a cut-fastball), inevitably, Detwiler got caught up in numbers games; games involving him being one of a number of capable pitchers vying for the fifth spot in the starting rotation and also those revolving around Det and his particular talents.

Talents that made him a first-rounder in the 2007 MLB Draft and led him to a breakout 1.6 fWAR 2012 season. Talents that have frustratingly been hampered by injury and a concomitant drop in fastball velocity, which, along with a lack of a reliable secondary pitch, has made Det a bit of a one-trick pony, something that is difficult to maintain in the starting rotation. As his reliever cohorts and former starting prospects Tyler Clippard, Rafael Soriano, and Craig Stammen can attest, the need for more than one effective pitch (and in reality, more than two) is just the tip of the iceberg when it comes to survival in the rotation.

Let’s look at what Detwiler did as a starter last season in comparison to starters with 150 or more innings pitched to see how far off he was in terms of providing effective innings as a back-of-the-rotation starter. We use the 150 inning criteria with the assumption that this number is roughly what is required from a number five starter.

First, the repertoire. Per FanGraphs and PITCHf/x data, Detwiler threw four pitches: a four-seam fastball (FA), a sinker (SI), a slider (SL), and a changeup (CH). Below, we see how often he threw those pitches in comparison to the NL average for pitchers with 150-plus innings, with standard deviations (SD) included to see how Detwiler slots:

NL Avg±SD 36±17.4 32.0±12.9 18.1±9.3 12.0±7.1
Detwiler 51.1 36.7 8.2 3.6

When summing up his fastball and sinker percentages into one value, we find that Detwiler threw hard stuff 87.8% of the time last year; when doing the same math for the average NL starters, we find the closest pitcher to that percentage is St. Louis Cardinal Shelby Miller, who threw some form of a fastball 73.6% of the time. Not surprisingly, Detwiler fell well below the average for secondary pitch percentages as well.

Now that the pitch frequency caveat has been beaten to death, let’s shift gears a bit and discuss how good a pitch was. Again turning to FanGraphs, we can use pitch type linear weights to help determine how successful a pitcher was with a given pitch. Below, we have Detwiler’s pitch type linear weights per 100 pitch basis (hence the ‘/C’ notation) for his 2013 offerings in comparison to the aforementioned NL starters with more than 150 IP:

NL Avg±SD 0.19±1.05 -0.32±0.39 0.06±2.07 0.04±1.53
Detwiler -0.12 -0.65 -3.55 -4.23

While me must be careful with how we interpret Detwiler’s offspeed results due to sample size and the potential for misclassification, we find that his fastballs (FA and SI) were both slightly below average, but still were within a standard deviation our NL averages. With the additional information the standard deviations give us, we find Det’s heaters were roughly average. We are also inclined to think that Detwiler’s secondary offerings were penalized somewhat by not only throwing so few of them, but also not being successful pitches when thrown in the form of high run expectancies. Simply put, even when he threw a secondary pitch, it more often than not led to a less than desirable outcome, be it a walk or a run scored.

So far, we have Detwiler relying heavily upon a pitch that was average and not throwing much else. Where did pitches end up once hit and was it any different than our average NL starter population?

NL Avg±SD 21.6±1.8 45.7±5.2 32.8±5.1 9.7±2.2
Detwiler 23.0 45.6 31.5 6.4

In general, Detwiler profiles as a starter here and more encouragingly, fares better than most in keeping homers at bay. The only red flag would be his line drive rate, which is borderline high.

Taking one step back in the process, let’s now look at where in the zone Detwiler pitches and how well hitters make contact, on average:

O-Swing% Z-Swing% O-Contact% Z-Contact% Zone% SwStr%
NL Avg±SD 30.5±2.7 62.5±2.7 63.3±6.3 88.2±2.5 49.5±3.5 9.3±1.7
Detwiler 32.2 66.9 75.6 92 50.7 6.6

For the above, ‘O-‘ stands for outside the zone, while ‘Z-‘ stands for inside the strike zone; SwStr% is swinging strike rate. By the looks of it, hitters swing a little more at pitches outside the zone on Detwiler. Contact rates against Detwiler are above the NL average in and out of the zone and at the same time, Detwiler does poorly with regards to getting hitters to swing and miss.

To paraphrase our findings, Detwiler is a one-pitch pitcher with said pitch being average in comparison to NL starters with at least 150 IP and his secondary offerings, while not seen often, are not up to par with other starters. With this, we also find Detwiler inducing higher than average amounts of contact, with below average swinging strike rates. While none of this completely dooms Detwiler and his chances of ever starting again, let’s discuss another piece of the starting pitcher puzzle that has escaped the lanky lefty, especially in 2013—success beyond the first time through an opponent’s batting order:

Time through order Detwiler wOBA NL Avg wOBA*
1st PA 0.302 0.309
2nd PA 0.393 0.317
3rd PA 0.413 0.336
4th+ PA 0** 0.310

*averages for all NL starters, no IP criteria used

**based on 2 PA

Using weighted on base average (wOBA) as our metric for the amount of offense against a pitcher, we find that NL starters tend to get hit a little more by the time they see a hitter a third time. Detwiler also displays this tendency, but in a more dramatic fashion, showing a 100 point swing in wOBA from the first and third plate appearance (PA) and a nearly 80 point swing in the third PA, in comparison to average NL starters. Here lies most of the damning evidence towards explaining sending Detwiler to the bullpen.

While these averages are from a 2013 season shortened by injury, the trends still stand; using Detwiler at his face value—a one-pitch pitcher—in shorter appearances more frequently lend him to be more productive and effective. Add to his difficulties in maintaining his health or consistent command of a second and third pitch and we see where the decision made shows the dedication that Nationals have in making the most of Ross Detwiler’s talents—talents that will serve him and the team well from the bullpen…for now.


Data courtesy of and


Stuart Wallace is a Contributor to District Sports Page. A neuroscientist by day, the Nevada native also moonlights as an Associate Managing Editor for Beyond the Box Score, stats intern at Baseball Prospectus, and a contributor at Camden Depot. A former pitcher, his brief career is sadly highlighted by giving up a lot of home runs to former National Johnny Estrada. You can follow him on Twitter @TClippardsSpecs.

About Stuart Wallace

Stuart Wallace is a Contributor to District Sports Page. A neuroscientist by day, the Nevada native also moonlights as an Associate Managing Editor for Beyond the Box Score, stats intern at Baseball Prospectus, and a contributor at Camden Depot. A former pitcher, his brief career is sadly highlighted by giving up a lot of home runs to former National Johnny Estrada. You can follow Stu on Twitter @TClippardsSpecs.


  1. […] Statistically Speaking: Detwiler falls short […]

%d bloggers like this: