At the bare minimum, Mr Cake appears to believe it.
He explains that his "great great great great grandfather" was a Unicorn, so there's a chance his family carried some surprising recessive genes, but then he says that Mrs Cake's "great aunt second cousin twice removed" (which makes no sense) was a Unicorn, so she carried some surprising recessive genes too, and then he does this odd thing with his eyes and asks "And that makes sense, right?"
He's not convinced, and he's afraid that she cheated on him, but he's gone into denial about it.
hmmm. it occurs to me that the fact that the aunt is the closest unicorn relative proves that Dirtpony genes are dominant over unicorn genes.
since the aunt is not a direct relative to mrs. cake, mrs. cakes great grand relative [who had the aunt as a sister] must have been D/u. if the great relative was u/u it would make no sense to use the aunt as an example because the great relative would be closer related to mrs. cake than the aunt, so the great relative must
not have been a unicorn, if the relative was d/d or d/p or p/p, there would be no unicorn genes in mrs. cakes family.
buuuuut, if dirt pony genes are dominant and mr. cake was the source of the pegasus genes, where did the second unicorn allele come from?
mrs. cake is D/u
mr. cake is D/p
shortcake is u/u, mr. cake cannot be shortcakes father if the closest living unicorn relative was an aunt.
tl;dr
if great aunt is the closest unicorn relative
great relative [who the aunt was a sister to] must not have been a unicorn.
if mrs. cake has unicon genes, they must have passed through great relative.
therefore great relative is D/u or P/u
either way, u is recessive.
if u is recessive, mr. cake could not have fathered both children. a p/u pony could though.