So if you want to know today's BMS estimate of battery capacity just charge to 100% and compare the available range to the rated range. For a new 3P that should be around 307-315 miles, possibly less if you've done something to the tires or aerodynamics.
Yes, this is what the OP should do. It's easiest. Tesla keeps it super simple and doesn't hide anything. It should be 315 miles, no less, for a new P.
And it will NOT be affected by tire selection or aerodynamics.
315 miles indicates a capacity of at least ~80.6kWh. That value includes the buffer, so translates to usable capacity of 95.5% of that, which is 77kWh.
This stuff is very deterministic. The only thing that is not deterministic is the BMS estimate, which is the very best estimate in the world of your pack's capacity. This estimate is reflected in the rated range at 100%, which should be compared to 315 rated miles in the case of this Model 3 Performance.
Here’s a screenshot of Tessie too
I can confidently say that their data is completely wrong.
I honestly can't make sense of any of their numbers. I can't make anything fit with what we concretely know to be exactly correct about this pack.
78.8kWh - your current pack capacity? Or what they think you started with (which would definitely be wrong as discussed)?
72.1kWh -> your "usable"
91.6% of the original
307 rated miles
78.8kWh your pack capacity? (As discussed this is not where your pack started, so that interpretation would be wrong.)
307 rated miles => 307/315 * 80.6kWh = 78.5kWh (2.5% loss, could be completely reasonable)
91.6% of the original would be 0.916*80.6kWh = 73.8kWh. Where did 91.6% come from?
72.1kWh "usable" => 72.1kWh/0.955 = 75.5kWh. WTF?
So literally nothing makes sense. I'd check to make sure Tessie is configured correctly and knows which vehicle you have. And again, the most useful thing is just to extrapolate your 100% charge value (or charge to 100% and see what it is).
Again, the usable is always 95.5% of the full pack capacity. That is rock solid true for any Model 3 except the LFPs. It's just always the case.
It's probably possible to translate the incorrect Tessie numbers though a mapping function, once you understand how they are reporting it incorrectly, going forward. Good times!