Here is an overlay of all the pitches I will be talking about in this post. Many thanks to Michael Augustine for making this beautiful overlay on short notice.
I am teaching a graduate measurements course this semester and the course requires a final project. I’ve taken to pointing out to our students that a modern cell phone is one hell of an instrument. Brandon Furman has a really slick phone (a Samsung Note 20) and wanted to see how useful it is as a measurement instrument. I suggested that he film some pitches in the lab and to find the ball trajectory. The results presented here are based on his video but are based on my code since this assignment isn’t due yet.
The setup is one I’ve used before but I will repeat it here. You can find more details here.
In our case, Cishek is replaced with the WSU cannon which releases 60″ in front of the camera. I have a Matlab script that reads the video and locates the ball (given an initial guess) and tracks it to the final location. It converts the location in the images to x, y, z, t using a drag model to estimate changes in velocity. It requires knowing the camera height, the location of the ground at home plate, and the initial velocity in addition to the number of pixels/foot at home.
The code also estimates the vertical acceleration of the ball. For pitches at 3:00 and 9:00, this should be just a little less that -1g (less than due to drag as the ball falls like a stone).
To give you a better idea of what the code does, if you use this video:
The code removes the background (which works better sometimes than others since the camera is jostled by the launch) and tracks the ball like this:
Here are computed trajectories for a fastball, changeup and curve.
We made 7 pitches with 2 baseballs. One was a 2019 MLB ball (29/1000″ seams). The other was a Wilson 1030 (50/1000″ seams). These are all flying a few feet short of a professional pitching distance, so the break is a bit reduced. The pitches are conventional fastball, curve and changeup (pitches that rely on Magnus force) and then a few loopers. You can read about loopers here. One thing I want to clear up: we throw loopers in the lab because we cannot throw other seam shifted wake pitches. If I could throw Jared Hughe’s sinker, I would.
The 9:00 spin on the MLB Loopers was a mistake. I intended to do 3:00. Since we are interested in vertical break here, it doesn’t matter a lot. Ignore the fact that those two break left instead of right.
Pitch 1 is a “Changeup” oriented so the seams don’t matter.
All Magnus force in this case is to the right. The ball is falling like a stone. As shown in the table, it falls at -0.95 g, meaning it has a little drag as it falls.
Pitch 2 is a looper with the loop on top. You should be able to freeze the playback and see the looper orientation. This pitch has the same speed, RPM and axis as Pitch 1, but you can see in the table that its acceleration is -1.13 g. It’s being pushed downward.
Pitch 3 is a looper with the loop on bottom.
This pitch has the same speed, RPM and axis as Pitch 1 and 2. It has a similar acceleration to the non SSW pitch, so I may have messed up its alignment in the cannon.
To give you and idea of how these accelerations compare to a pure backspin fastball and a pure topspin curveball, we include those also. These are -0.45g and -1.55g, respectively. The looper effect is small but significant compared to those large Magnus accelerations. Note that for Henry Aaron to be correct that fastballs rise, that number needs to be positive on fastball.
Here is the fastball, Pitch 4
And pitch 5, the Curve
And finally here’s the Wilson loopers. First, Loop on top.
And Wilson ball, Loop on bottom
Now, ending where we started, here are all these pitches, this time color coded corresponding to the table.
Wow, that video is fun to watch. There is only one of each pitch, but here is what I see: First, the fastball and curve break in opposite directions as we would hope, with gravity pulling on both of them down.
The 3:00 Changeup breaks to the right several inches. Recall, this is an MLB ball. Note that the MLB loopers (which are 9:00) move a similar distance to the left. The bottom loop arrives above the 3:00 changeup and the top loop lands below it, as advertised.
But the Wilson ball…. The difference between the top and bottom loop is similar to the MLB ball, but the Magnus effect is significantly diminished. If it were not, they would move to the right as much as the (blue) changeup. I’m not sure this is proof with only 1 shot of each, but it appears that the big seams on this ball kill Magnus effect in the looper orientation. The green and red ball should be as far to the right as the blue ball. They’ve barely broke right at all.
In conclusion:
- The Washington State cannon is amazing
- Augustine’s overlays are amazing
- Magnus effect is amazing
- Seam Shifted Wake can move a ball, but seam height may have a huge effect on what happens to Magnus. I am confident that some day, this result will make sense to me.