We ended up with 5 peer reviewers and an Editor override for one of our papers, instead of the usual 2 (I was the PhD student who actually did the algebra, so I saw all the comments.)

In fact, even after so much work to get it published, we now know of two errors in it - neither were pointed out by the reviewers of course.

I can't help thinking that that one experience sums up all of your problem cases above :)