Monday, April 24, 2017
Statcast Break on pitches
?One of things we take for granted is when we talk about the break on a pitch. We report this in terms of number of feet. This value is dependent on the starting point of the measurement. We can show this using the Alan Nathan Trajectory calculator, which Alan so generously has offered the community for a good ten years. The blue curved line (path of the pitched ball) and the red dotted line are from Alan's calculator, while the rest was added by me. The dotted red line represents the path of the ball with no spin. In other words, it's the tangent of the path of the ball (out of hand in this case).
And we can see that the gap between the straight red line and the curve blue line is about 1.7 feet. Since we can pick a tangent from any point of the curved line, we can pick one out at 42 feet out. When we draw that line (in green), we see the spinless ball would land at about 0.7 feet, or an offset of 1.0 feet with the spin.
So, did the ball break 1.7 feet, or 1.0 feet? Well, both. It broke 1.7 feet 55 feet out, and 1.0 feet 42 feet out. So in this illustration we have a scale factor of 1.7/1.0 or 1.7, depending on where you start the measurement.
?
Indeed, Alan noted that the amount of movement is proportional to the square of the distance it travels. So picking up the ball at 55 feet out or 42 feet out means that we'd expect a multiplier effect of 55/42 squared, or 1.71.
We can see this in the data being provided based on out of pitcher's hand (just over 54.4 feet from the back of home plate, or right around 53 feet from the front of home plate) using the 2017 standards in Statcast from the current vendor, compared to the prior standards of a previous vendor of around ~40 feet from the front of home plate. If we assume that that prior standard was 41 feet, then 53/41 squared is 1.67. And this is how it compares in practice, using the data provided by the current vendor, both based on current standards and prior standards:
?
So, why make the change? Well, as noted a few weeks ago, when discussing pitch speeds, we actually observe the pitch speed out of the hand. We observe the release point, out of the hand. Everything is based out-of-hand. Since we can measure things out of hand, we report the values being measured. So, movement is based out-of-hand. That's our starting point for the pitch data.
***
A special thank you to Alan for, well, basically, everything. The entirety of this post would not have been possible without Alan's generosity.