Thanks for the constuctive responses.
QAVION
On any level C or D simulator, systems simulation should be perfect. On today’s simulators the simulator manufacturers tend to run the same code as the aircraft computers, albeit on a different hardware platform. As a consequence the simulation should behave identically to the aircraft. Even on older simulators, there is no excuse for less than perfect systems simulation, although this requires far greater attention to detail by both the design engineer and the acceptance pilot to ensure simulator characteristics are as aircraft.
As Synthetic stated, Level D performance characteristics are solely dependant upon the quality/availability of accurate data. Today, there are, in my humble opinion only 2 areas of civil simulation that are not possible to replicate, namely sustained G forces and physical pressurisation cues. All other aspects are just a matter of cost as the technology is available.
Cornish Jack
The engine incident you refer to is in some ways a similar situation to unusual attitude training. No airframe manufacturer will or possible can provide real flight test data and therefore the airframe manufacturer provide engineering test data for the sim manufactures in conjunction with their customers to programme what they believe to be a realistic/representative scenario. Although in this specific case one would hope that the FDR information has been utilised to enhance catastrophic engine failure simulation. (UAL has a particularly well respected sim engineering group and I would imagine that they have already addressed this. Perhaps a UAL –400 pilot would comment)
There are two another engine related area of simulation that I think are particularly poor. On current Level D turboprop simulators propeller aerodynamic simulation, especially during pitch changes and engine failures tends to be unrepresentative. And on all simulators, slip ball simulation is basically cr*p. Again these are airframe manufacturers data issue.