Aside from the other reasons that this is a stupid idea; is anyone else baffled by how the proportions are supposed to work out?
I wasn’t able to find exact square footage numbers(and given the tendency toward unnecessarily high ceilings in status constructions; you’d really want the volume); but both the pictures of the house on fire, the street-level shots from one of those property-aggregator sites; and the aerial view from Google’s omniscient god-eye suggest that this house was not small.
What volume of ‘hot coals in the basement’ was he throwing around to have a plausible shot at forcing snakes out? If he has a snake problem to begin with it suggests that there is some porosity in the structure, which both gives the snakes opportunity to temporarily retreat and provides enough gas flow(especially of hot gas looking to rise) to scuttle the “actually, you can alter the atmosphere of a closed system quite dramatically, if you are patient” case.
And, unless snakes have some sort of natural, extreme, aversion to low levels of basic organic combustion byproducts(seems unlikely, given how many areas have an ecology that includes periodic fires; and panicked fleeing is only a good strategy if the energy cost and risk of predation are lower than the risk of being roasted to death); you’d need to achieve either enough smoke that smoke inhalation proves lethal even right next to the floor, in the snek segment of the room, or enough CO2 to kill by oxygen displacement, or enough CO to kill by hemoglobin binding.
Especially given that snakes take a fairly chill approach to metabolism and respiration compared to humans, that’s not a small amount of gasses required; especially since ‘burning coals’ isn’t the sort of well controlled environment that would let you coax everything to CO, which is probably the most efficient.
Just to get a ballpark comparison; apparently sulfuryl fluoride fumigation (‘tenting’) requires 6-16 ounces per thousand cubic feet, depending on the target, the temperature, and how long the building can be left sealed. With a house of this sort unlikely to be skirting the lower bounds with 7 foot ceilings(8 or 9, potentially more in some areas, seems more plausible) and what looks like a fairly large house we are talking a nontrivial amount of agent(though, very conveniently, the quoted LC50 for that fumigant and carbon monoxide (~1000ppm for moderately long exposures) are pretty similar; so it’s a better source of ballpark number than I’d originally anticipated).