I'm embarassed to bring this up again, because I seem to quote steven0461 too often--but, in something close to his words; "When you have eliminated the impossible, whatever remains is likely more improbable than an error in one of your impossibility proofs."