DrMAL: Included for completeness after Kasparov vs Korchnoi, 1983 on DVD #4 (of 5) "My Story" this is the first and stronger of two victories over Smyslov playing sharp 6...Qa5 (Cambridge Springs), in the latter game Kasparov vs Smyslov, 1984 Smyslov tried a new move 9.e5 instead. He won game 47 playing 6...Qa5 as black in Karpov vs Kasparov, 1985 adding a quip that any opening was good at that point. Kasparov commented on much sharper and complicated 7.cxd5 being best (he played more popular 7.Nd2 in later rounds for draws) stating with 7.Nd2 white has some advantage but "black has to equalize." It's interesting to me how Houdini clearly prefers 7.Nd2 over 7.cxd5 showing how little insight engines provide for opening choices.
He claimed 10.Bd3 as a novelty but it had been played a few times before including Alekhine vs Bogoljubov, 1929 a WC game. In fact, this one departs from that game by 10...e5 (instead of 10...h6 11.Bh4 e5). As such it seems 11...exd4 was odd compared to 11...h6 and it does not evaluate by Houdini as well either. "Classical player" Smyslov played this to give white an isolated queen pawn which Kasparov could care less about.
Since 12...h6 was now a mistake (13.Bxh6!) Smyslov made up for it with 12...f6?! that Kasparov called "a nice move" citing some merits such as covering e5. Aside from weakening Kasparov noted one demerit of exchanging bishop for knight (Nimzo drawback) which 14.a3 put the question to. From a classical view black has a nice game because of white's pawn structure where B vs. N does not matter as much in that thinking. 12...N7b6 was played in W Schmidt vs N Khechen, 1984 it was probably better.
Kasparov did not mention anything about 16...Be6 this move was evaluated as inaccurate by Houdini (compared with 16...Bg4 or 16...Re8 or 16...Ng6) with 17.Qb2! as even stronger than 17.Rfe1 this seems correct because 17.Qb2 keeps queens on while allowing c4 as well. Kasparov could have played 18.Qb2 probably strongest here but he chose 18.c4?! right away allowing the trade I don't think this was nearly as good, it loses all white's initiative.
20...Ne6 was surely best giving black an equal game. Kasparov said 20...Na4 was very good, wrongly criticized before. It was probably second best and good indeed for the reasons he pointed out, white's advantage was very small. He spent some time explaining how 21...b6 was better than 21...Rd7, true, but unimportant. Here his arguments differ with Houdini's evaluations, he stated 22...Rc8 was better than 22...Ne6 and 24...Nxc6 was a mistake compared with 24...bxc6 but neither really mattered.
Black's problems actually started with 26...Re8?! after 28.Bd6 white had a solid advantage. Then 29...Bd5?! instead of 29...Rd8 allowed 30.Rb8! to 31.Rf8 to which Smyslov replied 31...Ke6? instead of 31...Kg6 that should have lost to 32.Ba6! With 33.g3?! g6?! he had the same opportunity again and took it with 33.Ba6! Kasparov's finally stated at the end, this match was "easy." His analysis was mainly patronizing in the video and not usual for him, he glossed over this game because he did not care much about it. As he stated he would have been better off with Smyslov first and Korchnoi last now he had to go to Karpov.