All the recovered data pointed to the intentional approach that I described.
Such data is surely "evidence" of their intentions.
Let's make it really simple:
if you wanted to swing around the light house you need to cross the shoreline at a particular spot before it because you have a significant turning radius at speed;
at their speed, when crossing at waypointA (the LZ they had used before) on a track of 035 (that they were on on that last leg and which was on the HP's HoSI course selector) a banked turn of anything between 30 and 60 gives you plenty of room and (at their height) well clear of the mist running up the slope - BUT you would not want to get that crossing point wrong or come over it unexpectedly soon.
The use of something like CPLS would be the only thing that could explain why they would try it unless they were reckless.
These pilots were very good, unlikely in the extreme to be reckless, and were familiar with that area and low flying in the region and would not have trusted either TANS nor their eyes in those conditions to judge a turn that close in at any kind of speed.
The engine and control dispositions established at impact only fit with sudden awareness and reaction - neither a/c fault nor planned climbs, etc fit - they were surprised at where they were exactly.