If you wanted to do this experiment with reasonable statistics, it's quite amenable to parallelization. Pitch is cheap, funnels are cheap, dusty basement shelve

If you wanted to do this experiment with reasonable statistics, it's quite amenable to parallelization. Pitch is cheap, funnels are cheap, dusty basement shelves are (potentially) cheap, and webcams are inexpensive. Doing ~300 of these in parallel wouldn't be hard, and would allow considerable investigation of systematic effects. Decade-scale experiments aren't hard, but they do require planning.