Most aging Li-Ion batteries won't burn, we have seen that in laptops and smartphones for a long time now. What is the biggest problem today that leads to EV fire is NOT due to aging and dendrites in cells, they are design or manufacturing mistakes around battery or the pack.
Statistics though, I agree with you that it can be misleading. To use statistic correctly we need to have a large sample and a distribution that make sense (i.e. normal, bell curve) or it is just misleading to say all EVs are the same and you have 1/N chance of getting a fire. Imagine we have people here who said the average vehicle on the road is 12 years old, and conclude that all cars fail at year 24 and all the cars on the road have similar distribution across all year, then conclude that they are all about the same reliability wise and you can buy a 20 year old car and dump it on year 23 month 11.
Tesla fire or other EV fire are likely due to manufacturing failure, the Chevy are investigated and found to be pack design. Can we conclude every model and year being the same? or should we find out which one has which design and recall only a range of them? I am sure you already know the answer, and it is not statistic based.