KP, see above, # 15.
The scenarios which you discuss involve situational surprise. If a situation can be conceived and replicated in simulation then it is very difficult to be assured of creating fundamental surprise - and anyway it is only a simulation - the crew are primed.
Re pitot blocking; a classic example of inappropriate focus on a singe accident, the probability of which exceeded the assumptions of certification.
The regulators were fundamentally surprised, their ‘black swan’ event turned out to be a ‘muddy duck’; although a situation which could be conceived, it was discarded by the certification process in accepting triple redundant systems as having sufficient integrity.
Forget the code - modify the pitots, which was done; and furthermore, learning of the relative importance of speed, another system was added. Belt and braces; and triple redundancy is still good enough.