Wow. No. You want to use the mean of a sample. Certainly not the ideal. If you are going to fudge with numbers, then you instead should take the minimum to show what the worst case driver might expect.
Of course. However, you misunderstood my explanation. All I was saying is that the 57% reduction was not based on the average, it was based on the ideal. Blame AAA for using it as their basis, not me.
I'm sure they did it that way in order to get a bigger headline, so that they could quote a larger number as their “percentage range loss due to cold.”