Why do you say balancing is a myth? Every two or three months my 90D will lose a couple of miles that is easily recalibrated by driving it down to 10-20% and then charging up to 93-100%. This was all very well documented in wk057's research. It seems to be more a software algorithm thing in the BMS than a battery issue.
Balancing does take place, of course that is not a myth. I have been able to monitor my battery closely with TM-Spy for over a year now and was able to see how far the cells drift apart under what condition.
Partially charging does not cause the cells to drift apart. Partially charging causes the errors of the battery gauge to add up and it becomes less accurate. This is something Tesla officially stated in an email. The algorithm just gets less accurate and estimates on the lower end to be save. It looks like you are loosing range. Discharging to near 0% and charging to full (or almost full) recalibrated the gauge algorithm and you get the lost range back. Tesla explained it this way, but some people, for some unknown reason, prefer to say the cells get out of balance and then charging to 100 would re-balance them.
Looking at how much and when the modules in my car drift apart or not showed me that they actually do not drift apart much at all. Even when I partially charged for weeks. Jason noticed that balancing starts when 93% battery level is reach, but then it runs independent of state of charge or weather the car is plugged in or not. The bleeding resistors are very small so it could take many hours, maybe even days.