I certainly concur with His Dudeness that engine failures on take-off were 'beaten' into to us. Even in a two crew environment the pre-takeoff brief included considering the course of action if we lost one before and after v1. So I would not be surprised that is what the pilot thought was happening. It seems reasonable that he didn't abort the take-off at the early stages because the asymmetry would have presented itself gradually. And once he was airborne where the asymmetry was more clearly defined he may have waited for the auto-feather to do its thing only to realise, too late, it wasn't going to happen. I am not type-rated on the B200, but it was common practice in other aircraft type to go through the 'dead-leg dead engine' procedure. If he had done that I imagine that he might have been extremely surprised to discover that the yaw only got worse if he reduced the power on the 'dead' engine. At that point I can see how confusion would set in. Perhaps he reduced power on the other engine as well in an attempt to reduce the apparently inexplicable yaw?
Why are we speculating? Because we don't have the recordings that would tell us what actually happened. I read through the ATSB report and in the absence of those recordings the findings seem quite reasonable. The only thing I was disappointed with was that the ATSB didn't try throwing an out of rudder trim scenario in the simulator at a bunch of experienced, but naïve B200 pilots and see how they responded. Anyway, I am sure that scenario will become part of the future turbo-prop training regime.