Friday, 24 May 2019

The 2019 MyGolfSpy Ball Test

This first of a kind test has produced some surprising results and takes us in an unexpected direction. Prior to this test much of the opinion (mine included) was that a premium ball is a premium ball. The ball was effectively maxed out and there was little to differentiate between premium balls besides feel.

I had previously noted a difference in ball speed between the Chrome Soft and harder balls (such as the ProV1x and TP5x. The ST cannot account for aerodynamic differences and these appear to compund the results achieved by the Chrome Soft (& x) producing disappointing data for the ball.

However, what really got my attention was the variability MGS noted between balls/shots. We as golfers hope that if we execute the ball will find its target (assuming no signficant wind). The fact they noted significant variability when a Robot was hitting suggests this may not be the case.

MGS provided data for the mean and standard deviation for each test/ball combination. Going back to my Monte Carlo simulator I used this data to plot potential dispersion circles for the balls and the results are quite shocking:
You can see from this plot the MTBX is longest and Chrome Soft shortest. What lept out at me was the dispersion: If you play a hypothetical fairway of 40 yards (+- 20 yards of the mean dispersion) you can then count how many shots would miss the fairway:

This cannot be correct, a robot using the MTB X ball would miss the fairway 33% of the time!?!?!


  1. Can you suggest a ball for a swing speed between 75 and 80 mph. Is this where the low compression balls are longer ?

  2. I agree with you that this can't be correct. Been asking MGS since day one to tell us how many balls of each brand they hit for each swing speed. They will not say and that worries me that it was only a few. They also never tested the accuracy of the robot by hitting the same exact golf ball over and over to get a baseline dispersion for comparison. Without doing that, we really have no idea what the dispersion is for each brand. Maybe ProV1 is best, but maybe it just got lucky and the few balls the robot hit all landed together. We don't know.

  3. Thanks for plotting the data, MGS really opened a can of worms and Pandora's box on golf balls. There are a lot of interesting data, but other data is strange. Your graphs can easily show something is wrong, or just unexplained.

  4. Some data in the MGS review I also find highly strange, in that the Srixon Q Tour had excellent dispersion with Wedge and 7 Iron but bad one with the Driver. Also back in their test webpage they do not comment any of the comments there. I guess their test is still quite interesting but it is far from answering all the questions.

  5. James, love your blog. Please check out my Advanced Analytics for Skytrak at, you might like it. Also: do you know the formula for converting backspin and sidespin numbers to spin axis, and vice-versa (total spin and spin axis to backspin and sidespin)? The SkyTrak database stores only backspin and sidespin, no spin axis, so I'd like to calculate spin axis for my Advanced Analytics. If you know the formula, please email me at ronhornbaker at gmail dot com. Thanks!

    1. Looks like some interesting visualizations, I will take a look.

      I got help with the calculations from someone on gsf, they are actually pretty simple:

      Back Spin = Total Spin * cos (Spin Axis)
      Side Spin = Total Spin * sin (Spin Axis)

      Spin Axis = atan (Side Spin / Back Spin)

      Total Spin = Back Spin / cos (Spin Axis)
      Total Spin = Side Spin / sin (Spin Axis

    2. Thank you James! I'm currently using spinaxis = (sidespin/backspin) * 45 which I've seen others use and gets it close, but I'll switch over to the formula you provided later this morning.

      To try out Advanced Analytics, just sign up at then look for the Analytics menu item, and follow the instructions from there.

    3. James, just fyi using Javascript, the formula of Spin Axis = atan (Side Spin / Back Spin) doesn't work because it returns the result in radians. Converting radians to degrees using this as a final solution works perfectly, thanks again!

      spinaxis = Math.atan(sidespin/backspin) * (180/Math.PI)

      Plugging in equal numbers for sidespin and backspin gives the expected result of 45, and all other results look good as well. Updating Advanced Analytics now.

  6. James, shouldn't Total Spin = Back Spin + Side Spin? Because in SkyTrak, that is not the case.

    1. No I dont think so. The ball only spins around a single axis which is tilted when there is curvature. Side spin is almost a made up value to aid in visualization?

      The trackman university has good info on spin axis

    2. Thanks James, I just learned in my post at that Total Spin = Math.sqrt(Math.pow(Backspin,2) + Math.pow(Sidespin,2)), so I've updated my Advanced Analytics accordingly.

      Also got confirmation from Seth (VP Dev at SkyTrak) that the SkyTrak hardware appears to have a low threshold of total spin measurement of 2024 for some reason, which is something I hope they fix soon.