thanks for clarifying - it's fair to say that such extrapolations go beyond the premise of the question, and i think raising them is a bit of a misdirection. the fairy may provoke surprising questions, but, as you say, that goes both ways (is the fairy the only one that exists? how did the walrus get there?), and these aren't necessary to answer to evaluate the surprisingness of the observable situation at hand.
the point that i think people are really getting hung up on is that, by your original understanding and beliefs, a fairy being at your door was impossible, while a walrus would be (however extremely unlikely) possible. your argument (i suspect intentionally for the bit, which i admire) seems to miss the forest for the trees by claiming the walrus is more surprising because its unlikelihood is additive, requiring multiple assumptions about the world to be wrong, compared to the fairy's one. by contrast, it seems like many people consider the unlikelihoods weighted and multiplicative, akin to the drake equation:
(possible factors that could lead a fairy to be at your door) * (your estimation of the probability that fairies exist) = 0
(possible factors that could lead a walrus to be at your door) * (your estimation of the probability that walruses exist) = 1/(your desired extremely high number)
it's not necessary to refute the entirety of human reason to contrast these probabilities, as they both remain within the observer's perspective - what's more surprising, something you thought was impossible, or something you thought was extremely unlikely? i acknowledge that individuals might differ on this answer philosophically, but that's my best case for what i think the essence of the counterargument is. beyond that, i worry i may not be up to the dialectical tusk.
