They don't come much more sceptical than me but I went to a fire, quite a few years ago now, that we concluded started due to sunlight being focussed by a thick glass ornament onto curtains. When we looked at the evidence, there was absolutely no alternative cause of fire (no electrics, no smokers, no one in the room, no motives, etc.), the pattern of damage was completely consistent with the fire having started in the curtains, the ornament was still there still focussing the rays of sunlight to a point where the curtains would have been hanging, the curtain material was considered to be ignitable by this source.
As Sherlock would have it, "How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?"
And think of this. If the probability of this happening is, for example, one in a million (in a given period) then we only need glass ornaments on one million south facing window sills (out of the 30 million households in the UK) to make the occurrence what we would call 'expected' within the given period. That is to say, we should be surprised if it doesn't happen!
Stu
