Brian, I am aware of the limitations of simulators in that respect; indeed it's a whole separate topic in itself.
Obviously, no-one goes out flying and fails tail rotors deliberately to gain absolute data to put in a simulator programme! What is programmed into a simulator when insufficient data is available is a best guess by the programmer and is known as "off-model" simulation.
Despite this limitation, what simulator training can do is to make the pilot more aware of the different scenarios he might one day be faced with in the air. Before the RAF began using helicopter simulators (Puma was first, using the Helikopter Service facility in Stavanger), the information given was insufficient and incorrect. Certainly during my initial training, we only practiced the stuck pedal situation and engine offs, as per the t/r drive failure scenario. MOD thankfully realised this was inadequate because there are so many different variations on the tail rotor failure theme and made provision to expand the relevant training. There is some reference to this in the CAA document I have linked to; albeit the main mention is about the Army Lynx.