<<If you refer to the engine remains as showing stable paired power output, again not a record other than the state at the moment of impact.>>
If you check it out, there was more to it - some historical data was obtained from one of the engine computers (whatever they're called) and no emergency power demands had been recorded - and the engines were closely matched to a degree which is normally arrived at by a steady state for a period in which either the pilot via his beep trim or the FADEC trims them. The final manoeuvre included a full pull up of the thrust lever/collective - the obvious conclusion from the engine data and this control position is that the time between the start of this manoeuvre and impact was so short that the engine system had not had time to react and so disturb the balanced state - they were taken by surprise - they had not been struggling with controls before this final, urgent action which would almost certainly have left the engine records and state rather different to what they were.
Why is it so many of you try so hard to argue against any and every piece of data having any value or relevance to the crash - some data is surely good? Still the old mantra of nothing can be known? Take your blinkers off just for an exercise - while you argue each point severally, you are missing the point that so many parameters correlate to a particular result very strongly. I would go so far as to say that everything that is known about this crash points to an intentional approach to that LZ with a misjudged closing range resulting in an overshoot.
Why they were going there is unknown and likely causes for that causal misjudgement are pure speculation - but the evidence for that failed approach is strong.