You're probably over thinking this. Michigan doesn't really yield protect all that much. We were actually over-enrolled this year because the yield was too high. Also, from experience LSN isn't entirely accurate. It's self-reported data and a large number of applicants have never heard of it. From talking to people it seems like those with really high LSATs last year did way less research than those with lower scores and didn't post their data.
So they may under admit to guard against over enrolling this year then. (same effect as YP).