You may laugh, but this particular one is about calibrating the slingshot. When you shoot it normally, you have something called a 'draw' - a length to which the rubber bands extend before the pouch is being released. For experienced shooters, and I'm hardly one of them, the draw is quite constant and consistent.
One way of estimating the initial velocity of the ball bearing ball used as ammunition is to shoot it vertically and using a little complicated model, which is a combination of SUVAT equations modified by two different version of air drag (up and down - two because the initial velocity when shooting up is usually higher than the terminal velocity of the ball coming down). Then by using this model and the time of the return of the ball back to the surface (you can hear it striking the ground when lucky), you can estimate the initial velocity of the ball, and thus, its kinetic energy and momentum.
The problem is that when shooting vertically, the draw length is significantly different from the normal horizontal draw, and also, it is difficult to measure in either case or keep the same. One idea was to double the rubber bands with strings (I was going to use a thin 1 mm radio dial cord, which I believe is made of Dacron, but it doesn't stretch much anyway), tie the strings so that they would 'record' the length of the horizontal draw, and then lock them so that the draw is the same (i.e. this time the locked strings will limit the vertical draw to the same length as horizontal thus keeping the elastic energy stored in the extended rubber bands about the same as in horizontal shooting, thus resulting in about the same initial velocity of the ball).
The best I could think of was [1019] or a double version of it, which kind of works, but it tends to jam, so I need to cut it and redo it for each trial.
So yes, it is not a life-critical application (and probably would not inspire as much interest, controversy, and emotions as e.g. the Scott's Bowline saga I recently followed :-), but I was hoping that someone would know a better, less jamming option. If you think that [1019] can not be really improved further, that would be quite useful too as I will stop looking.
Cheers!