I'd generally assume that the base efficiency numbers behind the EPA scores are all calculated fairly and accurately, and I doubt that's why the Tesla miles seem to be shorter than the Porsche miles...
But after you calculate the efficiency numbers, there's a "adjustment" number you can calculate or apply. This is intended to account for times when you'll get worse efficiency because it's cold or it's windy or it's dark or it's wet or you're driving too fast, etc, etc. You can choose to either use 70% or to calculate it out, and you pick which method you use by which car.
Most everyone applies an adjustment of 0.7 (claim only 70% of the mileage you calculated)... But not Tesla. Tesla wants credit (reasonably?) for their heat pump and other things they've done to make sure the efficiency is consistent during different situations, and so they calculate their adjustment...and get ~75% [1]. So that's about a 5% "shorter mile" in ideal conditions. Tesla also sets you run the battery to basically dead, but then hides ~5% of that battery under "0 miles remaining/0%" to get you to not use it, while still letting it count to the advertised range.[2]
Combine those together, and you get tesla effective ("I'm not gonna hit 0% willingly, and it's a good sunny day, so the efficiency gains in the cold don't help me/hurt other cars") ~10% shorter mile[3].
I just wish numbers were more comparable and more accurate. More accurate would require the EPA to change the highway speed from ~56 (per above) to ~75. More comparable would be to get everyone to use the same 5-cycle methodology, and have a more useful/driver-centric definition of "empty"
[1]
The Secret Adjustment Factor Tesla Uses to Get Its Big EPA Range Numbers .... I think it's higher now than when that story was originally written?
[2]
Tesla saves more range at 0% than other EVs, test finds (Also an Edmunds story)
[3] The C&D feature at [1] above guesses the battery element is more like 20%. Testing at [2] didn't show THAT big a bottom-buffer past 0, but my guess is that the bottom buffer is used to also hide battery degradation. This would imply an larger actual difference in "mile length".