There was possibly a bit of wind at the end. It would take a mere breeze to effect the results. Also, the difference at the top is so small that decrease/increase is probably within the margin of error. 3wheeler, a very smart engineering type, has had problems with coast down repeatability.
It would also be interesting to know the span of time over which the data was gathered. Looks like it would have taken several hours, and I suspect tire temperatures would have changed within that timeframe. There is another subtile question in the data. One has a tendancy to read the data top to bottom, BUT, the data was taken the reverse order with the high pressure data taken first. I kinda wonder if the tires actually warmed up a bit during the test cycle due to rubber hysteresis?
I'm going to repeat the test on an Insight when we get a bit warmer spring weather, but I'm going to do it by driving my long test course. That would be more representative than the cold stuff we are having right now. Looks like MetroMPG gathered his data on a warm day.
In any case we known that the improvement with increasing pressure is a diminishing effect. At some point the improvement has to get too small to measure, though I have some doubt that it actually goes negative.
