pf, #7. Perhaps you should reconsider your expectations. Simulators are only models of the real aircraft; “all models are wrong, but some are useful” - the aptly named George Box, FRS.
A conclusion of ‘rubbish’ relates more to an expectation opposed to the real aircraft. Unfortunately many people use simulators inappropriately because of this, inferring performance or manoeuvre beyond the data used in the model.
733 #9, re ‘sensations’; these are central components of a simulated replication of an aircraft by providing a feedback loop for control - consider the overall system, man, machine, and environment (real vs simulated). A good representation of sensation can mask a poorly simulated aircraft. More often the reverse, particularly with reduced side force / acceleration where a poor representations of force or visual scene conflict with the aircraft dynamics; this mental conflict creates additional workload or need for physical compensation. Also consider how an expert pilot views a simulation vs a novice pilot, the latter might adapt by forming a technique for the sim which also applies to the aircraft ... or not !
Directional control with nose wheel steering might involve larger side forces or visual yaw depiction than that used in flight, thus it is of lesser quality, possibly limited by sim motion - how much sim tilt is required to replicate side force, how accurate, how long (for acceleration). Also, force feedback is important. Does the NWS tiller provide accurate force representation vs that which might be provided by rudder steering - relate this to the simulated environment, a dry runway vs simulated wet/slippery vs crosswind. (The last old sim I flew never met tyhe book landing distance on a wet runway - negative training for a landing overrun, ... or a positive contribution to emerg evac drills after a ‘surprise’).
Overall I agree with the concerns, and also with continued growth and emphasis on simulator training the machines (models) must have adequate fidelity for the task - and even greater fidelity for learning from mistakes, perhaps something beyond the training objective. Do we best learn from doing it ‘right’ or after a mistake?
I suspect that many regulators, having previously specified the quality of simulators, now envisage using them beyond their current capabilities with the risk of negative training.
Cost is always a driver, thus use a sim opposed to an aircraft; but perhaps the industry is approaching the boundary where using an aircraft will provide the greater cost benefit, in particular for those experiences and levels of confidence which cannot be simulated.
‘Simulators can create surprise, but they might never create fear potential startle’.