There are (at least) two further constraints on final location under reasonable assumptions.
For now I think that we can assume that the plane did not stop to refuel in the early hours of 8/3 (but we may reconsider this later).
One is the flying time of less than 6 hours from near Penang to the final resting place (but presumably not much less).
The other is the fuel which would give a similar upper bound.
Thus we just need to draw another circle on the map of radius about 5000km centred on the last radar blip (off Penang).
Then we intersect circles and arcs and should get just one or two possible approximate locations.
You ought to add one more constraint: erase any portion of the shaded circles and arcs which would require crossing a known radar range. While we can't be sure the radar operator was alert and inquisitive on the night in question, we can be sure that data will be reviewed and analyzed the same way the Malaysian radar data was shared with investigators.
I personally would put my next paycheck on that analysis already being done by now.