[Criminal Number 18F]

The thing is, the error was only half the controllers. Yes, he told the Russian plane to do the wrong thing. But the Russian pilots acted on what the controller told them to do, and not on the contrary commands coming from their TCAS.

In that instance, you are supposed to ignore the man and obey the machine -- the machine is (as in this case) less fallible. But the natural human reaction is to trust the human more.

This sounds like it is tailor made to be a CRM scenario. Can you do interactive ATC with a live human in the Frasca sims, or do they only give you a canned result? What about the MD? (I am assuming that the CRJ sim is still too ate-up to be trying anything but a handful of canned scenarios still).

In fact, it would also be a very, very good, publishable research experiment. Put crews in the box, and give them an escalating situation culminating in contradictory TCAS and ATC inputs. My hypothesis is that in most cases, crews not specifically trained for this scenario will trust the human over the machine.

d.o.l.

Criminal Number 18F

[Archangelsk]

Unfortunately, the Russians were trained to listen to the controller. An unfortunate validation of Pavlov.