I wonder if those who seem to believe AFCIs have made people safer from electrical fires can offer up any kind of verified statistical proof that they have reduced the number of house fires since their introduction in (I believe) the 1999 NEC. Or even any proof that they work as claimed by the manufacturer.
Even if the AFCI can do every thing we are told it can do and even if prevented 100% of the dwelling unit fires that are said to be of electrical origin, it would be a long time before the number of fires would be statistically significant.
There are ~131,7 million dwelling units in the US.
There are ~370,000 dwelling unit fires per year the last few years.( note, that is about 50% of the number of dwelling unit fires that occurred in 1980...there was a steady drop between 1980 and 2000, with very small drops since then)
Of those 370,000 fires about 48,100 of them are said to be of electrical origin. (13% of the total dwelling unit fires)
That is a rate of ~365 dwelling unit fires per million housing units per year
~85% of the dwelling unit fires that are said to be of electrical origin are in dwelling units that are over 20 years old
That is a rate of ~55 dwelling unit electrical fires per million new dwelling units per year
There were ~1,080,000 new housing units built in the US in 2014
Assuming that the AFCI would prevent 100% of the electrical fires in all of the new housing units built in 2014 you could expect to prevent ~59 fires that first year of full compliance with the 2014 code. (note the 2014 code does not require AFCI protection of all dwelling unit circuits, so the actual number of fires prevented would be less than 59)