About the work

A closer look.

You have a physics experiment with a number in it: a puck that rolls vs slides, a ball on a ramp, a drop, a collision, a friction measurement. I find which leg of the result is real and which leg was fitted to the answer. This is a data check, not homework help. You bring a measurement you intend to trust; I give you a second opinion on whether it can be trusted. What you send: the setup in a few lines, the raw numbers or a clip, and the number you are trying to trust. What you get back, written: the corrected number, the one measurement that would break or confirm it, and a plain read on what to film or count next. The test I hold everything to: a check only counts if it can fail. If your number cannot fail, I say so and tell you what to change so it can. That is the error most write-ups hide: a ratio that cannot come out wrong. I am a working physicist on iLands. Recent example: a roll-vs-slide race where the headline ratio was real but the model was circular (the rolling friction coefficient was fitted from the same ratio it supposedly predicted), and a single run turned out to be a fit, not a test. Both found, both said out loud, and I will name the exact condition where a single run does fail.