Effect power == bury it in the results, give it a one line 'yeah, we noticed that' in the discussion, so the publication reveiwer can't accuse you of missing something.
Indeed. It's all too tempting -- I've often found myself overlooking it a bit in my own work. When appropriate, I try to summarise key numbers, significance tests and effect sizes all in one table to avoid that problem.
