After the AF447 accident I ran a scenario at the end of a two day recurrent if there was enough time remaining. Clearly stated as no jeopardy; let's take a look at something that will be of benefit; but not saying what it would be. The failure I gave was left and right static port blockages at 50' after takeoff. In ten events only two crews identified the problem prior to heading for a serious overspeed or crash. Of course I froze the sim prior to any negative event occurring and pointed out the problem. Once the problem was known all ten crews were easily able to get the airplane back around for an ILS to a landing using the Airspeed Unreliable checklist.
The point here is this: Any sim training where the problem is known in advance is lacking in reality. Almost every crew can easily get the airplane back on the ground once they know what the problem is. The accidents that are occurring are because the crew is not recognizing the problem in time to take corrective action.
The more modern the airplane the more serious the issues when pitot or static systems are out since all of the computers are relying on this information. This results in some seriously misleading cues such as stick shakers; stall warning; overspeed warnings; poor flight control feel; etc.
Obviously any training and awareness given as to what pitot/static/AoA faults look like is good, but the best training is the surprise effect in order to help crews learn to recognize the fault in a timely manner.