Probabilities
It's important not to look at unconditional probabilities.
For example, it is highly unlikely that a specific flight will end up the way MH370 did. And yet, MH370 had the extremely unlikely result.
If we say that it's highly unlikely that cause X could happen, that doesn't help. What's important is how likely is cause X compared to cause Y, under the condition that the MH370 event actually happened.
Thus cause X can have a very small probability. But given that this event happened, the probability in this event is obviously much higher. If we summed up all of the unlikely causes, the result would be 1 (100%). The important probability is whether X caused this accident, out of the set of all possible causes. So, 10 scenarios (for example), each with an unconditional probability of less than 1/1,000,000 add up to 100% when you take this into account, because under this condition, they would average 10% (1/10) rather than 1/1,000,000.
The key here is to look at the relative probabilities: rank cause X and cause Y etc and compare them only with each other.
SAR officials understand this. Conditional probabilities of this sort (and with lots of other things thrown in) are fundamental to assigning SAR assets - picking what grids to search and when and how and how often to search them, adjusting as the conditional probabilities change in response to the results of negative findings.