You’re right, Spoffy. I’d misread it as one pair of letter pairs, as it were, for two solutions, rather than two different pairs of letter pairs for four. But, yes, there’s still the shared first-half pairs. I suppose it’s the different perspectives of starting with solutions then working backwards, as a setter might, and starting with the clues then working forwards, as a solver. He must have seen the second-half letter-pairs as more of an issue than we do (my head’s hurting now...)