If it was you, wouldn't you want to see if the firmware solves the issue before doing something like that?
It depends. If I had the resources (like a good friend with a flatbed truck and a great set of backroads to drive), I wouldn't think twice about doing it for the data's sake. But if I was going to be forced to pay full towing fees downtown in a city, I would have to think twice. That's assuming I really care - and you do care, if you aren't willing to accept Tesla's statement at face value.
Until then, we have no data on which to make any conclusions whatsoever - including the primary conclusion people in this thread are jumping to, which is that the 90 kWh packs are losing real-world range because of the indication on the dash. This worry could all be hand-wringing. To some people, it's going to be a worry that affects them, to some it doesn't.
You can't just discharge to 3-4%, because you don't know if there's a 30 mile floor underneath that. You literally have to take it down to the shutdown voltage and then get it back on a charger immediately. Only then will you know if you really lost that range.
Right now, based on what I'm seeing from 85 kWh packs, I wouldn't reach any conclusions whatsoever until April/May, when the peak rated range is achieved for a given SOC level. It's dangerous to take a comparison of cold month vs. warm month and try to reach any conclusions.
- - - Updated - - -
Heck, given wk057's recent posts on his hacking thread that the 85 kWh battery isn't really 85 kWh anyway I don't see the value in running it down to 0 not knowing what number I should be expecting.
That's why you have to compare the product: Your Wh/mi * your miles should roughly equal rated Wh/mi * rated miles. If it's significantly less, then you know that you suffered. But you have to do it in one continuous drive, so you don't lose power on vampire draw, while car's in park, etc.
- - - Updated - - -
For me at least; seasonality, temperature whatever you want to call it has 0% to do with my displayed rated range. I live in a temperate climate and the range has not fluctuated very much between last November and now. Having said that I trust and respect the data I've seen from @FlasherZ and @MarcG which does seem to show a difference for them.
It's unfortunate, but the only direct comparisons you can really do will have to be at the one-year period, if you want to exclude any effect you might be seeing from it.
If you haven't signed up for the TeslaLogger, you might want to consider that as it will capture this data for you so that you can graph it easily.