Robson Lowe has this to say about the pence issues (in regard to overprints on German New Guinea, which they appear to adopt for the Marshall Islands stamps, too):
On the pence values, each row was printed in turn from a "set-up" of ten subjects which differ from each other in greater or lesser degree; on later settings, the "set-up" consisted of only five subjects.
To produce the pence values, there were at least nine settings, some of which were used only on a single value; after the 5th setting the remaining "set-ups" held only five subjects.
The original German sheets for the pfennig values were in sheets of 100 stamps, printed 10x10. It makes sense that they would have made up one horizontal row of overprint settings, ten subjects wide, and then overprinted each row, one at a time.
I take from that description that vertical pairs are probably going to be identical and horizontal pairs are likely to be different "in greater or lesser degree."