SkyTrak’s algorithm

[DannyMRICS]

Just looking at some of my data, and noticed something interesting (well, to a golf nerd!):

I had two shots, side-by-side, as follows, both good shots with not much side spin (ball speed, LA, backspin, sidespin, carry):
1) 117mph / 16.9* / 6501 / -93 / 162y
2) 117mph / 17.8* / 5978 / -428 / 167y

5y difference with very similar numbers. I put the same numbers into Flightscope and got carry numbers of 168.6y and 169.8y, respectively - only 1y difference.

The difference between flight algorithms is well documented, and that isn’t the point of this post. What I thought was interesting was the sensitivity of Skytrak’s Algorithm to minor changes in LA / backspin. I’d noticed a higher degree of sensitivity to driver LA and spin than I would have thought too.

I hit balls on the range the other day (a mis-match of balls) and one ball ‘carried’ 165y with 2,300 backspin; another 142y with 7,800 backspin, both with a ball speed of 106mph and LAs within 2*.

Any thoughts?

Dan

[Clevited]

I like nerding about this stuff too, so you are not alone.

Just my loose observations regarding my ST, I can't help but feel high launch is rewarded a little more than it should be, and spin is more detrimental than it should be. High launch, low spin drivers for example carry further than Trajectory Optimizer would show I think, but low launch high spin would show less than Trajectory optimizer would.

If I hit a 7 iron with like 24 degrees of vertical launch and Skytrak misreads spin to be like 2500-3500, it will travel miles. When I hit 7 iron with 24 degrees and it sees the spin right (nearly 8k most of the time), it won't travel as far as my real life 7 iron does all weather condition settings the same (altitude, temp, no wind). There must be some thresholds where the math in the algorithm deviates a lot. Perhaps you have found the point at which that happens?

It sort of feels like the flight algo is a result of applying a trendline to data from another machine. Like taking tons of Trackman data and trying to create the algorithm off of that. There will be times it blows up and is wrong depending on how many points were recorded and the variety of points. I have done a similar thing and have a similar problem. I made a smash factor calculator that simply uses PGA tour average stats to calculate smash factor and therefore clubhead speed based on launch numbers Skytrak spits out. It blows up at times on me.

[DannyMRICS]

I tend to agree.

I used my Dad's driver the other day, and got quite a bit more carry with maybe 500rpm less backspin. My father got quite excited at such an easy extra few yards, but I was much more muted - I suspect there is very little difference in real life. I was looking at the Flightscope xi base model (about £1,000) which tracks the whole ball flight, and was thinking using the xi with a camera based system would almost give you a Trackman, without the club data.

[Supraracer]

Just looking at some of my data, and noticed something interesting (well, to a golf nerd!):

I hit balls on the range the other day (a mis-match of balls) and one ball ‘carried’ 165y with 2,300 backspin; another 142y with 7,800 backspin, both with a ball speed of 106mph and LAs within 2*.

Any thoughts?

Dan

Perhaps if it was outdoors, wind could have been a factor or lighting issues.

[DannyMRICS]

Sorry, the inverted commas were meant to imply that the carry was calculated by Skytrak - they weren't actual carry numbers.

[Dave Lee]

I like posting this guys site, even though seems like he doesn't update it anymore.. It has a ton of data comparing different monitors

https://exploringimpact.blogspot.com/2017/03/efficiency-spin-axis.html

This one is just an example of similar data being treated pretty differently across the different algos. This other one where he moved the skytrak up and down 30 mm show's similar impact of spin and LA.

https://exploringimpact.blogspot.com/2017/01/does-height-matter.html

I think the bottom line is you're probably going to have to do some self calibration if you want to directly translate to your real life game. Even with range balls and a trackman you still kind of have to adjust for your own ball and wind.

[DannyMRICS]

Yes, I really like his posts - very interesting.

I think the bottom line is you're probably going to have to do some self calibration if you want to directly translate to your real life game.

I completely agree - I intend to get on the course, hit with SkyTrak and laser my distances, and compare with Skytrak's distances to sort of self-calibrate.

My point was really more to do with the sudden change in calculated distance with only a slight change in LA and spin.

[LEO MODE]

Just looking at some of my data, and noticed something interesting (well, to a golf nerd!):

I had two shots, side-by-side, as follows, both good shots with not much side spin (ball speed, LA, backspin, sidespin, carry):
1) 117mph / 16.9* / 6501 / -93 / 162y
2) 117mph / 17.8* / 5978 / -428 / 167y

5y difference with very similar numbers. I put the same numbers into Flightscope and got carry numbers of 168.6y and 169.8y, respectively - only 1y difference.

The difference between flight algorithms is well documented, and that isn’t the point of this post. What I thought was interesting was the sensitivity of Skytrak’s Algorithm to minor changes in LA / backspin. I’d noticed a higher degree of sensitivity to driver LA and spin than I would have thought too.

I hit balls on the range the other day (a mis-match of balls) and one ball ‘carried’ 165y with 2,300 backspin; another 142y with 7,800 backspin, both with a ball speed of 106mph and LAs within 2*.

Any thoughts?

Dan

For me, I wouldn't recommend using Flightscope's algorithm because it tends to fluctuate carry distances a bit longer if the ball speed is fast (like 140+mph fast) and spin is low (like less than 2000rpm low). I figure Skytrak algorithm is fairly better representation of real life like Trackman outdoor and I personally use/compare that with OptimalFlight algorithm. You would notice carry being longer on GC2/GCQ/FSX algorithm as well.