Mr. Robot definitely factors in the 3% hit from the talents. Look at your BiB solution, as an example.
Sometimes the optimal set of gear ends up higher than the hit cap because of itemization. In this case it incidentally landed on 9% in the BiS solution, making it seem like an error.
We spent a lot of time making sure that hit rating was handled correctly in the classic code, as it is something that comes up all the time.
We even wrote up a few posts explaining how it all works, for reference: