In theory simulators are very accurate. In reality they might be. I'd expect the simulator for an existing airframe with decades of service (like your 747-400)

In theory simulators are very accurate. In reality they might be. I'd expect the simulator for an existing airframe with decades of service (like your 747-400) to be more accurate than that for a new airframe. The primary purpose of simulators is training, not testing. While both the airframe and its simulator are under development, substantial inaccuracies wouldn't surprise me. TFA quotes this dude's attorney thus: "The simulator was not reading right and had to be fixed to fly like the real plane." Is that true? Is that ass-covering? How are we going to find out from this dude's testimony? Other evidence must be examined. TFA says this pilot had a hard time with the simulator. We also know that these planes crash very often. We can't just put those two propositions together to prove that the simulator predicted these planes would often crash in this particular way.