The secret behind both Andrew Miller and Jake Diekman's success lies in two major mechanical or repertoire-related changes. Miller's breakout season was preceded by years of stagnation in the rotation, but a shift to the bullpen brought about the start of his improvement. Diekman, on the other hand, had been a reliever for almost his entire professional career (since 2009). The changes Diekman took, though, resemble the very ones that Miller took in order to advance his career.
Andrew Miller had once been considered a failed starting pitching prospect. Taken with the sixth overall pick in 2006, Miller never caught on because his control wavered. Through his first six years in the majors, he had conjured up a 5.4 BB/9. His lack of control held back his progress as a starter until the Red Sox finally realized he simply couldn't hold up in a rotation. His first year exclusively in relief was a drastic improvement because his strikeout rate rose to 11.4 K/9. Now that his role on the team was not in question, Miller could begin to take the crucial steps towards becoming the knockout relief ace he is today.
The first major step that worked for Miller was abandoning his tertiary pitches and solely focusing on his four-seam fastball and slider. As a starter, Miller had used a changeup and sinker in varying degrees during his first six years in the majors. Now primarily a two-pitch reliever, Miller's performance with the fastball and slider improved greatly. As can be seen from the graph below, Miller's reliance on his slider gradually increased over time and now rests at about the same rate as his fourseamer.
Jake Diekman's progress is derived from the same repertoire adjustments. A changeup was once another weapon in his arsenal, but he now only pitches with a slider and sinker. With more time to focus on those two pitches, they skyrocketed in value. Admittedly, it was only a small portion of his percentage pitch usage, but it still required time on his part off the field.
In my mind, the much more effective step towards becoming a great relief pitcher has to do with the stance Miller took on the mound. As a starter, he often leaned too far left (from the pitcher's POV) and as a result, lost control of his pitches. As a pitcher compacts their stance and throws from a more central location on the rubber, they're able to command their pitches much better. Gradually, Miller drew his delivery towards the center of the mound and was able to harness his repertoire and hit the strike zone where he desired. The graph below demonstrates the evolution of his horizontal release point.
Jake Diekman's horizontal release point shows the exact same pattern. In 2012, he was throwing from four feet to the left of the mound (from the pitcher's POV), but in 2014, Diekman is pitching from only two and a half feet to the left of the mound. In this case, 2.5 feet is a perfectly reasonable horizontal release point. The command and control followed suit, with his strikeout percentage rising from 26.7 percent to 32.0 percent, and his walk rate dropping from 15.3 percent to 11.2 percent.
With these changes to their horizontal release points, both pitchers were able to exhibit superior arsenals. To demonstrate this effect, the graphs below plot each southpaw's horizontal release point versus their pitch velocity. The most noticeable trend is that as a pitcher brings the release point closer to the center of the mound or their body (which is indistinguishable without looking at vertical release points or game film), the faster the pitch travels. Diekman's graph is much better at portraying the trend than Miller's, but Miller's still follows the general trend.
The proof is in the data. Jake Diekman's mechanical and repertoire-related change has made him into a formidable reliever, and Andrew Miller's changes finally wiped the failed prospect label away. Even though velocity and command are not the only factors when analyzing a pitcher, they are the building blocks of pitching and have to be weighted as such. If more pitchers embrace a slightly more compact stance or abandon their superfluous pitches, then they too can take the next step.
. . .