So, Jack, I have not been able to find out a whole lot, but here is what I got...
First, I did not find trace of an 1892 perpetual calendar patent. What I did find is that on May 23rd 1889, Patek was awarded patent N. 1018 for a pocket watch perpetual calendar mechanism, so I presume you are referring to this one.
This patent covered a QP mechanism which Jean Adrien Philippe, who died in 1894, had been using for some time in various iterations, which were all subsequent refinements of the previous variants.
My understanding is that the main objective of the patent, however, was not the compactness of the mechanism (those were, after all, pocket-watch days - the first wrist iteration of a QP by Patek only saw the light in 1925, and only entered regular production in the very early '40s), but rather flatness, and - primarily - the instantaneousness and simultaneousness of the jump of the various indicators.
In my opinion, using a 12-step program wheel, with the only variable step addressed by a specific sub-wheel, is a much more elegant and refined solution from a theoretical perspective than a 48-step program wheel, though the latter might be simpler to implement (and perhaps less prone to failure in the long run?). But not being a watchmaker I doubt that my opinion counts for a whole lot...
Alberto Schileo
Patek Philippe Forum Moderator
Tempus fugit, horologia manent