yes, this is exactly what my pack has gone thru... twice in 4 years actually. First time was when I loaned the car to a guy and he drove the pack until the car wasn't even 'limping' any more. Second time I was caught taking an unexpected side trip 'rural' and thought I could make it back to town before pack voltage got too low. Both times forced one cell to reverse polarity under load, and caused an internal short in the cell, rendering it very obviously dead.... if you top balance and drive your vehicle extreme discharge you risk that low imbalance and causing a cell to go empty before the others and kill a cell or two and even damage surrounding cells.
my concern is that *some* other cells are damaged as well, but may be usable for a while longer.
I KNOWINGLY accepted the risk of top-balancing because I want to allow my charger to work the way it it supposed to and bring the pack up to a 'nearly-full' state automatically, where hopefully all the cells finish at about the same time and none suffer from over voltage. BUT, now that my pack has suffered two catastrophic under-voltage events, I suspect the capacity and IR of several more cells are going to force me to replace some cells soon. I am trying to prolong that and hesitate to mix in new cells with old, but I'd also like to avoid the cost of an entirely new pack if half the cells are still usable and relatively undamaged.
My main purpose for this thread from the beginning was to try and QUANTIFY how bad is bad, and how to measure that so I can have some confidence in deciding how many cells I have to replace when the time comes. i.e. exactly what difference in IR or capacity will make particular cells 'unusable' with regard to drifting over multiple charge cycles (because of damage resulting in differing IR and/or capacity).
My conclusion at this point is that I need to review the capabilities of my Cellpro, and see if I can get any meaningful measurements matching up to which cells start 'drifting' too high at end-of-charge. In the meantime, I have bypassed dead cells, and lowered my setting for the pack end-of-charge voltage to hopefully prevent any single cells from overcharging in a normal charge cycle.