Ok, it seems too challenging, or maybe not challenging enough...
I doubt it be the second one, so to get it started, here is my progress so far:

In the situation described above, when the view is switched to the tank's (camera at the tank's location and looking the same way the tank is facing) the centre of the board can be defined as follows:
Wx = [ ( c + a ) / 2 + ( ( c - a ) / 2 ) cos(2RTx) ] sin[ angleUCX cos(RTx) + RTy] + [ (Ty) sin(RTx) ] cos(RTy)
Wz = - [ a cos(RTy) - b sin(RTy) ] cos(RTz) + (Ty) sin(RTx)
where
a:= BoardSize.x/2 - Tx;
b:= BoardSize.z/2 - Tz;
c:= sqrt( sqr(a) + sqr(b) );
angleUCX := arctan ( b/a );

I am pretty confident about the above equations, even though I cannot certify them 100%.
For the y coordinate, however, I could not work out a complete equation, but I did one only including RTx and RTy (no z rotation). It is as follows:
Wy = b sin(RTx) - (Ty) cos(RTx).

I have tested the previous for many combinations of rotations and displacements of the tank, and they seem to work.
What I still have not managed, is 1) to get a full equation for Wy including RTz; 2) equations for RWx, RWy and RWz to suit the situation I described in the previous post.

Hope this helps someone help me...!
Any math genius out there