I once had a book on the sharp Wilkes-Barre variation of the Two Knights Defense written in the 1970s in which there were a few moves which were said to be clearly winning for one side in which a computer analysis of the position said otherwise. The situations were unusual, for example the last position in the analysis would have three threats in it and the conclusion would be that one side is clearly winning. Then the computer would find a move responding to all three threats and say that the other side has the advantage.
So my question is 'What are some books on chess openings that show how a computer rates the top move choices after analyzing each for a certain amount of time?'