If I'm not wildly mistaken, Greg of Gregs garage has called it a day after "talking BMS" all these years. However Peter "retepsnikrep" should be around.
You are wildly mistaken,
Both myself and Peter are still around. I've been running my Matiz for a year now, testing the bms and Peter and I are slowly
working on a new slave board setup based on the pic 12F1822 which has an internal voltage reference, see the bms thread for details.
To answer wattsin's question, I am not aware of any catastrophic cell failures which weren't caused by an external factor, shorts, loose connection or overcharging. The main danger appears to be the electrolyte which is flammable. Anything that causes the cell to get too hot can set the electrolyte on fire.
But I am aware of quite a few non-catastrophic cell failures on packs that weren't being mistreated. Typical failure is very high internal resistance in the cell. This can happen at any time on a cell that is otherwise working fine, with no warning that a cell is on its way out, one minute it is working fine and the next minute it is dead. Failure rate can be as high as 2% or 1 in 50 cells. I recommend buying some spare cells when making a pack, just in case.