Your argument is based on a false analogy ... 'a dog in another case was wrong, therefore it is necessary to assume the dog in this may have been wrong'
We do not have to look at the set of circumstances most likely to 'fit' in with the cadaver dog's alerts
The dog's alerts are a factual piece of evidence and what those alerts indicated will not be proven ( or disproven ) by the imaginery scenarios you are presenting and asking others to comment on
Is this thread about EVRD dogs in general or about Eddie and Keela?
Are we not allowed to consider examples of other EVRD dogs in order to understand what issues may have affected the performance of these two dogs in particular?
Where have I stated that it is "necessary to assume" anything at all?
You don't have to do anything you don't want to, I thought however that if we took the EVRD dogs in this case to have been 100% reliable indicator that Madeleine died in the apartment then it would be interesting to discuss what that means in practical terms. Perhaps you will be more interested in discussing this on a separate thread so I shall start one.
I am not seeking to prove or disprove anything, I'm just having an informal chat on an internet forum.