“Probably arson, prove it wrong.”
That’s not the way things work. No one has to prove a negative.
Uhmmm. Disproving a theory is exactly how things are supposed to work.
A good theory includes the conditions which would disprove that theory if the specific data can be observed. That results in a pile of discarded theories and what we have left is considered "proven". Evidence which confirms a theory does not prove that theory is correct, it just makes us feel better.
"Probably Arson" is an assessment and is a testable theory. The test is simply too expensive for us to do. Kind of an unfair challenge method, and really more of a declaration of perception bias than anything else.
The 25% Arson rate assessment by NFPA is a strong contradiction to the "Probably Arson" theory" simply because 24% is less than 50%. Words have meanings after all.
Oh, wait, the theorist for "Probably Arson" really means "Probably Arson if the facility was a Food Processor". We do not have enough public information to disprove that assertion, but it could be obtained at considerable expense. The theory is testable and could be disproven.
Considering the consequences of the theory it might be a good idea to put a lot more effort into testing it.
I tailered the exception for you