Cottonwood
Roadster#433, Model S#S37
Since this got brought up in this thread again. After heading out to Mount Rainier again this past weekend I'm now 100% positive that the trip planner uses the 30 mile average to help try to adjust the predictions. Coming down the mountain it makes wildly optimistic predictions which once you put 30 miles behind all that regen, become more realistic. Obviously this adjustment factor is there to help adjust for driving style. But I'm surprised there isn't some sort of floor to avoid situations where driving down hill for long periods distorts it.
Interesting...
I've found the Energy:Trip display of predicted, percent SoC at the destination to be pretty good with elevation changes, up and down hills, etc. It seems to use a recent sample vs it's model to update the prediction for actual driving usage, but the sample seems to be 5-10 miles and not 30 miles (I do have my basic Energy Screen at 30 miles). There does seem to be some bias to overestimating the destination SoC while going downhill and underestimating the destination SoC while going uphill, but I have only seen a few percent change on big, Rocky Mountain grades.