The chosen answer is wrong, the right answer is "it depends". The whole point of the "For Update" mark on a cursor declaration is to ensire that ITU locks are taken early, not late; in cases where there is risk of deadlock, this minimises the both the frequency of deadlock and the amount of work that will be thrown away (rolled back) when deadlock occurs.

It's not surprising that it's easy to demonstrate that a cursor definition designed to optimise performace in cases where there is significant c=oncurrent access and risk of deadlock will not be optimal when it's the only query being run against the data concerned. It's also not useful to make that demonstration.