Green Dot
And for facts you need the data and evidence to prove a theory.
If evidence remains unavailable, the next best thing is try and simulate the conditions and observe systems behavior against the data available. If the cause is still unknown at this time, apparently something happened which cannot be retrieved from the recorded data.
Ah, a true scientist among us
However this kind of "let's see what happens" stuff has too many drawbacks.
You can never be sure that you have actually got all the varriables tinkered to perfection and thus any answer that you would like to draw is challengeable by all except yourself.
It's akin to running an experiement to see how it breaks. When all that we can really use is to confirm that it doesn't break enough times to be deemed safe.
In the long and short of it we can't build an airplane that won't break, all we can do is build one that won't break most of the time (there is a difference)