Throw (Good) Curveballs with Men on Base

Two weeks ago, I showed that pitchers can throw curveballs that reliably depress the exit velocities of hitters. There were many of directions I wanted to go with this, including determining whether the ability persists year to year (I believe it does, generally - a pitcher can always work on his curveball) and whether the ability presents itself in the minors (I imagine it would, but it's possible that pitchers really develop their curveball early at the MLB level rather than in the minors). Both pieces of follow-up research would offer strong opportunities for action at the Major League level; scouting and fairly shallow research could be used to identify pitchers with a top-notch second pitch early on. However, both pieces of follow-up research require something that doesn't exist: data. Baseball Savant doesn't have complete 2015 Statcast records, and no Statcast data is available for Minor League teams, at least publicly. If you're a professional baseball team that wants me to answer these questions...

Instead, I turned my attention to slices of data that did exist within the available Statcast data. Spin rate didn't show any correlation with the depression of exit velocity. This was a little surprising given what we know about spin rates and curveballs. But spin rate is not in and of itself a good thing; there can be meaningless spin on a ball, and curveballs are among the pitches that tend to have the largest amount of total spin being meaningless. Perhaps separating useful from meaningless spin on curveballs and comparing it to exit velocity depression would yield more valuable results.

I also looked at the bast state at the time of the pitch. considering any instance with at least one man on base to be an instance in which the pitcher would throw his pitch from the stretch. This is not always true; some pitchers pitch from the stretch in bases-empty situations, and some pitch from the windup regardless. Some mix it up! In general, my expectation was to find that pitchers would do worse from the stretch overall, whether by an inability to generate torque and spin with a faster move to the plate, or a push to reach back and throw harder curveballs that don't drop as much, and as such get hit for line drives more often.

The results were counter to my expectations: pitchers tended to depress exit velocities at the same rate with men on base than with the bases empty. Right handed pitchers shown because the Orioles only had righties throw curveballs in 2016:

It seems that, in general, pitchers that can throw good curveballs can throw them in any situation. Pitchers whose curveballs depress exit velocity do so consistently.

This is particularly useful for crafting pitch sequences with runners in scoring position. In some cases, fastballs are considered ideal to prevent the opportunity for runners to steal. However, with a pitcher on the mound whose curveball regularly depresses exit velocities (ie, induces ground balls), it may be advantageous to rely more heavily on the curve and ask the pitcher to generate outs rather than hold runners. At the very least, pitchers with good curveballs should not shy away from using them when the other team is threatening.