1) Starting the watch and the buggy at the same time, and then stopping them at the same time and;
2) whoever is driving that Tumble Buggy CAN'T DRIVE STRAIGHT!
We started using a countdown before starting the buggy, and then had the person with the watch stop the buggy. Still not 100% accurate, but we had to try and get it as close to accurate, and consistent as we could. But the other problem made us think. With the cart turning off to the side, should we try and keep it straight? Would that interfere with the results? Was the cart turning part of the trial, and should be treated as part of the normal motions? Well, if we touched the cart during it's timed runs, it left the thought about skewing the data. Because if it slows down, or speeds up, because it got bumped by us, or something we placed, that would change the outcome. So that wouldn't work. But if we left it, would it remain consistent? Could that be considered part of the cart's motion, and therefore be left in the data as a totally valid factor?
As it turned out, we calculated the cart distance by how far, in a straight line, it traveled. So even though it curved, we marked where it stopped, and measured straight from the start, to a spot which was in line with where it stopped. We felt that that was the most consistent way to handle the curve, without skewing the data.