Robert Gsellman’s Ominous Velocity and Spin Trends

Last year, Robert Gsellman came up to the big leagues and helped stabilize an injury-plagued Mets rotation. With a power sinker that had added velocity, the unheralded Gsellman missed bats and generated grounders at an elite level. And after another Steven Matz injury put the lefty’s dependability into question this past spring training, Gsellman was given a prime opportunity to grab a 2017 rotation spot and run with it. To date, however, he hasn’t returned to top form.

For one thing, the results have been ugly. By RA9-WAR, Gsellman has been one of MLB’s worst starters. That once-great sinker whiff rate has been halved. But beyond outcome-level stats, his pitch data indicates worsened stuff. Overall, Gsellman’s sinker is down 0.71 mph, and especially striking are his in-game velocity declines.

Below are LOESS-smoothed curves plotting the difference between the given two-seam fastball velocity and its initial “baseline” in that game — represented, in this case, by the average velocity of the pitcher’s first five two-seamers. By restating velocities like this, each start becomes its own “universe” and we mitigate pitch-tracking biases on the game and park level.

Out of the gate of his 2017 starts, Gsellman’s velocity is dropping. By the 40-pitch mark, he has typically lost 1 mph from his starting speed. As he approaches 80 pitches, his two-seamers are nearly 1.75 mph slower. The orange curve does rebound near its end, but a widened 95% confidence interval reflects a smaller sample of pitches and less certainty that he’ll continue to gain velocity back. Regardless, Gsellman is ending his starts at 1.5 mph off his opening speed. Compared to the dark gray curve for the league — which indicates starts this April in which pitchers threw 20-plus two-seamers/sinkers and 90-plus total pitches — Gsellman’s velocity has tumbled much more steeply.

In-game velocity drops were not such an issue last year. Gsellman’s velocity would actually rise slightly in the first few innings before falling under the baseline once he eclipsed the 30-pitch mark. Whereas Gsellman is spending the bulk of his 2017 starts below his initial baselines, his 2016 velocity descents were gentler and closely aligned with the MLB norm. We do see a steeper downward dive towards the end of the blue curve, but again, smaller pitch totals near the century mark widen the confidence band and limit our certainty.

Notably, Gsellman attributed last Wednesday’s in-game speed drop to tightness. But even if that start is removed, the persistent distance between the 2016 and 2017 confidence intervals shows a significant, worsened trend.

By way of illustration, here’s the same chart as above, but without the April 26 start:

A velocity dip is troubling, of course, for two big reasons. Not only would it portend a fall from his previously rosy outlook, but it may foreshadow an arm injury. Diminished pitch spin is also a major injury predictor, and the news for Gsellman isn’t good on that front, either.

Gsellman’s spin rates are down across every pitch type. His two least-used pitches — i.e. the four-seam fastball and changeup — have declined less than 30 revolutions per minute, declining by just a few percentile ranking slots. His two-seamer and curveball, however, have lost enough RPM to drop down by some 10 percentile slots. And most glaring of all, Gsellman’s slider has plunged by 172 RPM, enough to fall halfway down the spin rankings.

These trends in velocity and spin hint at an injured pitcher. At the very least, they’re producing a less effective pitcher. It’s not good timing for the Mets, who are now down two starters and have few enticing MLB-ready pitchers in the minors. And it’s not good timing for Gsellman’s pursuit of a lasting big-league rotation job, an opportunity he may be reluctant to pass up. But right now, he looks like a different pitcher, and returning to last year’s promising form will depend on his capacity to regain last year’s stuff.