I have a Model 3 LR, delivered during early 2020. It has about 30,000 miles and, to date, has been absolutely trouble free. Since the beginning, I have tended to keep the charge limit between 70-90% and very rarely let the charge state fall below 50%. My commute round trip is 19 miles with an altitude change of 400ft, so the average round trip uses approximately 6%. By April of this year, my calculated maximum range was 262 miles, which I thought was relatively poor, approaching SR+ territory. I pinged Tesla service, which responded, ultimately saying (paraphrased) "There is nothing you can do. It will continue to drop and when it reaches the threshold of 70% of rated range, we will take care of it". I did not believe him. Given the way I was charging the car and that I am a "light" driver (bordering on driving like an old fart), I doubted that I was looking at a hardware issue. I reasoned that the BMS has to use some combination of regression, a moving average, and God knows what else to come up with a value, with general order number one being "thou shalt not strand the driver by estimating more range than is available". So I decided to change my charging regime by simply using the middle 50%, running from 75% to 25% (ish) figuring that this would provide a wider range of data points for calculation. This meant not always plugging the car in once the state reached below 50%. With this regime, the calculated range steadily increased after that and is now at 296 - 299 miles. I did not expect that large a change figuring, if I was right, that it might go up as high as 285 or so, but it is still continuing to creep up. I believe the rated range a the time I purchased the car was about 321 (?) miles. If so, then the difference between the rating and the current estimate is 7-8%, well within the expected decrease that I would see at 3 years or so. It has been fun gaming the estimate to see if I could change it. I did not expect it to be so simple and effective. I have no idea how generally applicable this strategy might be.