Not All Shifted Wakes Are Equal: Post 61

Twitter: @jwillyg20

It’s time to define a Seam Shifted Wake (SSW) pitch more carefully. The main feature is that a baseball’s seams can cause earlier flow separation on one side of the ball relative to the other. This asymmetric wake is what we refer to as the “shifted” wake (see figure), and causes an unbalanced pressure distribution and a corresponding sideways force on the ball creating movement. This sideways force in addition to Magnus is the reason some systems cannot accurately predict the movement of pitches based on their spin axis and rpm, since they only consider Magnus, gravitational and drag forces.

In our early work, we noticed seams tended to cause this early separation near what we call the hemisphere plane of the ball. The hemisphere plane slices through the center of the ball and is 90 degrees to the trajectory of the ball. After analyzing many data images, we made the following “separation map.”

At the time of this map, we assumed seams located anywhere but the green area of the map didn’t have any effect on the location of the flow separation. It was also assumed, and built into our pitch trajectory simulator UMBA2, that if a seam is present in the green regions of the separation map, there will be a SSW force on the ball. As we have collected more data, it is clear this was over-simplified. There is still much to learn about the complex influence baseball seams have on flow separation. That is clearly demonstrated by these two images acquired this summer.

As you can see in the image on the left, the flow on top is not separating on the seam, but there is a shifted wake. In the image on the right, the separation occurs on the seam, but there is a symmetric wake and therefore no sideways force.

Here is my new claim: The seams further upstream play a vital role in where the flow separates from the ball. The wake on every pitch is determined by the location of all of the seams on the baseball.

In determining if a pitch has a shifted wake, it is important to not only analyze the location of the seam near the hemisphere plane, but also seams upstream, and the relative position of that seam. Additionally, not all shifted wakes are shifted to the same extent, meaning the strength of the SSW force on a pitch varies depending on the seam orientation.

Rather than automatically claiming a pitch has a shifted wake solely based on a seam being positioned within the active range of our old separation map, I’m now analyzing approximately 900 images to determine first if there is a shifted wake, and then the location of all of the seams that are playing a role in the creation of that shifted wake. We are seeing some interesting trends in the data that I look forward to sharing in the future.

So here’s the process I use to analyze the data. Note that all images shown in this post were acquired using Particle Image Velocimetry. If you are new to our measurements and how to interpret them, we have a primer here.

1. Look at the processed PIV image of the baseball. The separation point is typically where a velocity vector is shown normal to the surface of the ball, or where two vectors are pointing in opposite directions. The typical place to start looking is where the vorticity (the blue or red air) is especially dark. (Note: Ignore the answers inside the ball. The PIV sees the ball moving but all we care about is the air the ball is disturbing.)

2. After determining what I think is the separation location, I look at the raw data in the same location to refine the answer. There have been times where the processing or post-processing isn’t done quite right, so it’s important to analyze the raw data in order to confirm the separation location. The two raw images were taken microseconds apart from each other, so the fog particles in most of the image are not moving. However, the fog particles being disturbed by the baseball are moving, and it can be fairly easy to find the location where you can see the particles are moving vs. where they aren’t. This location is the separation point.

3. Center my on-screen protractor on the baseball

4. Record the angle from the hemisphere plane of the flow separations occurring on the ball. These balls are all moving horizontally to the left, so hemisphere plane is just a vertical slice through the ball. Additionally record the location of every seam on the ball.

This next part is what I am really excited about. As I stated previously, a shifted wake is not necessarily determined simply by a seam being located in the active region of the old separation map, and not all shifted wakes are equal. This led to the idea of a “shifted wake percentage” (SW%). The main idea is that there must be some maximum shifted wake possible, and every other shifted wake is a percentage of that one. The equation we came up with is this:

All of the angles used in the equation are reported as relative to the hemisphere plane, where positive angles indicate separation forward of the hemisphere plane and negative angles indicate separation behind the hemisphere plane. For the time being, I am assuming the most shifted wake possible would be if the flow separates on the top of the ball (0°) and at the back of the ball (-90°). After I determine the separation points for all the data sets, the maximum difference in top and bottom separation will be changed from 90° to the maximum difference found in the data sets. Some examples are shown here for clarification.

It should be noted that a positive SW% means the wake is being shifted up and a negative SW% means the wake is shifted down.

In my huge spreadsheet of all the data collected, the data are organized according to the location of the seam nearest the hemisphere plane on top of the ball. As I said above, I’m noticing some interesting trends of the SW% bouncing from negative to positive as the seam position moves around the hemisphere plane. The following plot comes from 288 images of baseballs in 4-seam orientations. Note the red dots shown are the orientations of the shifted and symmetric wakes shown previously.

This is where we are now. I’m currently analyzing about 900 images of data acquired in 4-seam orientations, 2-seam orientations, arbitrary orientations, and rotating data of each in order to help us make a new and improved pitch trajectory calculator. Can’t wait to share more results!