This is a ramble, probably not worth reading.
Fire starts, bummer, but what's going to happen to it. It's not going to spotaniosly spread, it needs the fuel right? Well, what fuel would allow spread. Smoke's a good one, but you need the smoke to be produced.
Anyway;
If you find an ignition source - a toaster or a pump. How many have been produced and how many have caught fire. In the case of chip pans (I stand to be corrected) there are 50 left in the country, how many catch fire? statistics say, 49 will catch fire of the 50 left. So 98% will give your probability (0.98). 10% of people have no brain, they will throw water on it and so spread it. (0.1). The house is concrete, so are the chairs and beds and dogs in it. So the fire spread potential is 0.1% (0.001). So fire is likely to happen and spread beyond the original seat. But once the original fuel is burnt, thats it, it'll go out.
Is that a high or low risk?
If the person wasn't there, a low risk - it's going to burn and thats it, won't go anywhere.
The person is stupid so they won't get anywhere, high risk?
But the risk to everyone else in the house, pretty low.
Saying that, they'll have terrible back from sleeping on concrete.
I think risk is too wide a subject to be able to define it further than low, medium and high. You can't. Unless you know 100% that something will occur with certain outcome, then it's not possible to say that it will or won't. This disturbs a lot of people. Uncertainty.
But, it has to be done. Certain things can be said for definate, they are what you must build a risk assessment on. Then you use science and engineering to back up what you have said.
Told you it wasn't worth reading.