As long as your errors are randomly distributed you should keep making observations until you get an error of -347% in the other direction, thus cancelling out your previous error and allowing you to model your entire existence with Riemann curves

My lab partner and I *