Endless spiral bug

[KevinMerryman]

I am getting a bug on my SkyTrak where the ball will go into a spiral and just go off into the distance forever. I have to completely close the program and relaunch it to stop the bug. I have only seen this off my driver so far and it happens about twice per practice session.

Anyone else seeing this? Is it a known bug?

Thanks

[Morini]

I have seen a about a dozen videos posted of this. It is believed to be quite rare. You are the first person I am aware of that has it happen regularly. I am sure seth is aware of it, and will fill us in. Clearly some type of loop in in the ball flight model under certain conditions. Maybe a rounding issue creating a zero denominator or something. Early on it was suggested that it was more likely if elevation and/or wind was jacked up. While you wait for a fix, maybe you could try moving the elevation a bit, as you seem to be creating the conditions for this to occur much more frequently than others.

[KevinMerryman]

Thanks Morini I was messing with the altitude a few weeks back so I will set that back down and see what happens.

[SkyTrak_Seth]

Yeah, it's something with the flight model algorithm not getting along with extreme weather/environment custom settings. I've only seen it with extremely unrealistic numbers though. I'll see if we can put a system in place to limit some of these settings so this doesn't happen at all. We've put some small checks in, but given the rarity it's been on the back burner.

[Dan McWhirter]

There's a lot of calculation to turn the pictures into numerical data. Transcendental numbers like e and pi, with no solvable exact value, are estimated to high accuracy with numerical expansion. Each term of this calculable expansion gives a more accurate value. For example:



If you calculate this expansion for an infinite number of terms, you may get the true value of e, or at least the best you could do with an infinity of calculations conducted for all time on a super computer. Notice the last term out there at infinity requires you calculate the factorial of infinity.
So you test the total summation through each term for the difference with the summation through the previous term, with the question: When the difference meets a target (converges to a desired accuracy), stop calculating.
Some expansions involve alternating + and - terms with more complexity. These can blow up on you, never reaching the desired convergence. I think that is when you get the calculated shot showing an endless spiral, because it can't be calculated!
But what could cause the solutions to be incalculable to within the desired accuracy? I suspect a sensor error which provides a nonsense value against all the other sensor values.

[Morini]

It usually happens when altitude is jacked way up. Probably causes a number to cross a point it was never expected to do. For example a divisor becoming less than one or a subtracting an unexpected negative.

[Dan McWhirter]

My first post since April...Must've gone into depression about then.