Recent experiments find large differences in binary blend miscibilities between poly(propylene) in its ordinary and head-to-head forms with poly(ethylene propylene). This miscibility variation cannot be understood from simple Flory-Huggins (FH) theory but can readily be explained with the lattice cluster theory (LCT), a generalization of FH theory that accounts for nonrandom mixing, monomer structure, and chain semiflexibility in a systematic manner. Calculations are also presented of miscibility trends for poly(propylene)/poly(ethylene butylene) blends, one of which has yet to be studied. Experiments for this system should provide a nontrivial test of the LCT predictions.