Agreed. In this situation, you have an observation that you saw an improvement in mpgs. We just don't know if it's the Ceratec (can be a coincidence) or related to something else or a combination of other variables. When I added a full-length belly pan to my car which will be aero on the undercarriage, my next tank had noticeably higher mpgs but it was short-lived but I could have easily stated it was from the belly pan. So in your dataset you see a jump up at the point of adding Ceratec that stayed consistent so basically a trend line with a kink in it?