One of the first challenges we encountered was a means to detect Seam Shifted Wake in pitches. Doing so requires a measurement of vertical and horizontal movement as well as spin metrics. Right now, if you are anyone other than a major league team (meaning you have a HawkEye system available), this requires a Trackman plus an independent means to measure spin. Spin could come from optical system like Rapsodo or Yakkertech or a “smart baseball” (PitchLogic or Diamond Kinetics). Just the Trackman alone is going to set you back north of $10K.
This post is about a solution to Trackman side of things, but it also has promise to deliver spin. Until then, spin can be had for $250. The main drawback to my scheme is that the results must be analyzed and are not available in the bullpen. This is not a small thing, but I believe the low cost more than offsets this drawback.
The basic idea is to have a high speed camera behind the pitcher at a known location (2D location is shown, but one must also know the distance to the side (z direction) of the camera.
Before I show how it works, let me demonstrate what it does. PitchTracker is based on high speed video acquired from behind the pitcher as shown below.
Based on the vertical and horizontal location of the ball in the video and using the initial velocity and a drag model to determine the ball’s y location at any point in time, the x(t), y(t) and z(t) of the ball is found. This information is everything required for the Level 1 metrics shown below. Here’s an example of x, y, and z for a few pitches.
It is easy to see that we obtain the release point in x and z, the initial and final angles and the pitch location.
PitchTracker output has 3 levels, with more/better equipment required for each successive level:
Level 1: Horizontal Break (HB), Vertical Break (VB), release points, release angles, approach angle, pitch location
Level 2: Axis Deviation, Non-Magnus movement
Level 3: Ball orientation relative to the axis.
Level 1 requires:
- A camera with at least 240 fps and a fixed focus
- A velocity measurement
Level 2 requires Level 1 plus
- A means to measure the 2D axis (i.e. Tilt) and gyro angle (i.e. Efficiency)
- A means to measure RPM
Level 3 will require
- At least 480 fps
- Sufficient zoom and depth of focus to view multiple rotations of the ball
- No Rapsodo/Smart ball required
- Some code development yet to be done
We have demonstrated Level 2 measurements to date. These measurements were performed by Spencer Davis at Next Level Athletic Development in Virginia. He used a Rapsodo and a Sony RX100 V camera. The pitcher is Brenan Hanifee. Results from two different bullpens are shown. The colored symbols are the average of a pitch time while the individual pitches are shown with a +.
We did not have an opportunity to make pitch by pitch comparison to Trackman, but I was able to compare a different pen with the same pitcher. The plot below comes from a Tread Athletics report on a pen Hanifee did there in between the two above. The break values comes from Trackman measurements. The two systems agree well.
In addition to break measurements, when coupled with spin data from Rapsodo or a smart ball, we can compare the “Inferred axis” to the “Observed axis” to find axis deviation or Axis Shift. An example for the sinkers from one pen is shown below, with axis deviation shown in degrees. Note that for a RHP, negative deviation on a pitch means in moved more arm side than its spin axis would suggest.
Similarly, one can compute non-Magnus movement by comparing the movement predicted by a Magnus model to the movement found from the video. For these same pitches, this is shown below in inches.
Bottom line, this is a poor man’s Trackman. And a Trackman plus a smart ball is a poor man’s HawkEye. I’m sure I’ve got them scared.
If you are interested in this service, contact me at bartonleesmith at gmail.com. I can guide you with the setup and discuss our pricing structure.