Does More Extension Yield Better Fastball Performance?

We try to make the connection between a pitcher's release point extension and how their fastball performs.

Earlier this month, Nick Piecoro wrote an article focusing on Bryce Jarvis and his attempts to improve his extension to make his fastball less hittable. More extension means the ball would have less distance to travel to home plate, allowing a fastball to get on the hitter quicker. Statcast measures it as Perceived Velocity, which is different from the velocity readings you see on T.V. or at the ballpark.

We're going to start with the hypothesis that better extension on the fastball can lead to better outcomes on the pitch. First we have to establish the relationship between extension and how its perceived velocity plays compared to the release speed.

The graph shows not only a positive relationship between extension, and perceived minus release speeds, but an r-squared value of 0.565. Additionally, every additional foot of extension at release adds 1.7 MPH to the perceived velocity reading of a pitcher's fastball. So there is some truth behind the prevailing wisdom that more extension allows a fastball to play better than its radar gun reading. 

With the concept of how extension affects a pitcher's perceived velocity established, the logical question is does more extension lead to better performance? The metric we're going to use for pitcher performance is wOBA, or weighted on-base average. In order to do this, we classified each pitcher by their perceived velocity from 82 to 99 MPH and took the average wOBA and xwOBA of all pitchers that fell into each velocity range. For example if a pitcher's perceived velocity was 94.2 or 94.8 MPH, it would be lumped into the 94 MPH grouping.

Perceived Velocity Table and wOBA/xwOBA at each velocity grouping

wOBA vs. Perceived Velocity Grouping: Bar Chart

The data once again shows the prevailing wisdom that more perceived velocity leads to better results. However, this only illustrates a general trend and not guaranteed to be a predictor of success. If we just take the 457 pitchers' perceived velocity and map it out against their wOBA, the general trend agrees with the bar chart, but has a very weak correlation overall with an r-squared of 0.032. 

wOBA vs. Perceived Velocity Scatterplot

With an r-squared of 0.032, that means only 3.2% of the variance of a pitcher's fastball performance, in terms of wOBA, can be attributed to its perceived velocity. That means we need to look at other aspects of a pitcher's fastball, and we need to look beyond just the release speed and extension in order to determine its effectiveness. Other factors such as spin rate, spin axis, percentage of active spin, usage, and location also likely play a role in determine how effective a fastball can be at getting hitters out.