I've never been involved in manufacturing batteries, but I was involved in a production engineering capacity in making limit switches that shut down oil fired residential furnaces if the burner didn't shutdown when it was supposed to. This was the last thing that would stop the furnace from starting a fire in a house.
I can assure you the scenario you describe above was not how we determined product quality. Any company who would use statistics in this manner for any high risk product would be out of business very quickly.
That doesn't mean mistakes never happen. It's almost impossible to cover every design, manufacturing and end user scenario 100% of the time, but large companies are quite risk averse and do everything they can to cover themselves against law suits by making sure their products are safe to the best of their ability.
The Dreamliner li-ion battery fires is a case where something was overlooked. Probably doesn't have anything to do with the cells themselves or I think someone would have found the problem. I suspect a lot of complexity in the electronic battery control system and a lot of other electronics close by too. AFAIK they still haven't figured out what happened, or been able to recreate the problem. I don't envy the people working on this one