In your latest analysis, after 53.Rc6+, I cannot find a way for Black to hold the position. However, the whole variation starting with 27...Nxd5, resulted in Black having some surprisingly good resources. It was much closer than what I expected when I first reviewed the position after 27...Nxd5.

In fact, even after 53.Rc6+, the position is very close to a well known position from Fine's, "Basic Chess Endings":

click for larger view

In this position White to play wins and Black to play draws.

In the position we reached after 53.Rc6+, note the similarities to Fine's position. However, I was unable to find a way to reach Fine's position with Black on the move.

click for larger view

.

Here is a 3rd position from the game Bernstein-Cukiermann 1929:

click for larger view

There followed: 1...Ra1 2.Ra7+ Kg8 3.g6?? (3.h6! wins) 3...a2 4.Kg2 (threatening 5.h6) Rb1 5.Rxa2 Rb5 6.Ra8+ Kg7 7.Ra7+ Kg8 8.Rh7 Rg5?? (Rb3! draws) 9.Kh3 and wins.

And a 4th position from the famous 1893 Tarrasch - Chigorin match:

click for larger view

In this game Chigorin selected the wrong plan and lost after 1...Ra2? 2.Kg4 Ra1 3.Ra6+. Tarrasch's notes to this game state that he believed the game was won even if Black played 1...a2. However, Chigorin later pointed out the correct defense: 1...a2! 2.h5+ Kf6 3.Kh4 Rh2+ 4.Kg4 Rb2 5.Ra6+ Kf7 6.Kg5 Rb5+! (Black dare not allow his King to be driven to the back rank) 7.Kh4 Rb2 8.g4 Rc2 (not 8...Kg7? 9.h6+ and 10.Kh5 winning) 9.h6 Rc6! and draws.

At move 28, 28.Rh3 was better than 28.Bf6+.

At move 29, 29.Rh3 was better than 29.Qb2+.

At move 30, 30.Qd4 was better than 30.Rh3.

At move 31, 31.Ne5+ was better than 31.Qd4.

At this point, White's advantage had been reduced to a minimum. Black could now continue (.13) (18 ply) (Fritz 9) with 31...Bc5 32.Ne5+ Kg8 33.Qd2 Qd6 or he could continue as he actually played; (.16) (19 ply) with 31...Kg8. Fritz indicates the best continuation after 31...Kg8, is 32.Nxd6 Qxd6 33.Rh5 Qg6.

Morphy chose 31...Kg8, and then Anderssen played 32.Rh6, instead of the stronger 32.Nxd6. Anderssen was now at a small disadvantage (-.45) (18 ply) 32...Bf8. Fritz gives the following line as the best continuation: 33.d6 Qb5 34.Ne3 Rd7 35.Rh5 Rxd6 37.Rg5+ Rg6 37.a4 Qc5.

Morphy's choice of 33...Rf7 put the game back to an equal basis (.00) (18 ply). Fritz indicates the best continuation is 34.Rh3 Qa4 35.Rg3+ Bg7.

The game continued 34.Rh3 Qa4. Then Anderssen again made a choice not recommended as best by Fritz, for his 35th move. I will review the game's final 8 moves later.

Beginning at move 27, Bxf6 starts figuring in on every move. Karsten Mueller on the Morphy CD gives 27 Bxf6 as an alternative to 27 Nc4 which has the practical defect of inviting the mind numbing complications of "27...Nxd5 28 Rb5!"

(BTW this 1 move line is the extent to which they analyse a critical juncture, and I have to ask, why do chess authors such as Knaak and Mueller give such abbreviated lines, when they have all the space in the world. The lines are not in the least obvious, and could have made their CD immensely more valuable.)

On Anderssen's 35th move he could have taken a draw by 35.Rg3+ Bg7 36.Rxg7+ Rxg7 37.Qd5+ Kh8 38.Ne5 Rf8 39.d7 Rd8 40.Qe6 Rg7xd7 41.Nxd7 Qxd7 42.Qf6+.

Instead, Anderssen played 35.Rc1?. This move gave Morphy a great opportunity by playing 35...h6! 36.Rg3+ Bg7 37.h3 Kh7 38.Qd5 Qd7 39.Rc2 Rc5 40.Qd1 Qe6 41.Rd2 Rd7. After 35.Rc1?, Fritz evaluated the position as (-.90) (20 ply), but Black's advantage increases during this variation. I believe it is a winning line for Black.

Morphy on his move then played 35...Rc5. Sergeant comments, <This stops the threat Qd5 - Qe6.> However, Fritz rates 35...Rb5 as equal (.00) (18 ply), Morphy had just missed his winning opportunity. Fritz supplies the following continuation: 35...Rc5 36.Rg3+ Bg7 37.h3 h6 38.Qd2 Kh7 39.Ne5 Qa5 40.Rgc3 Bxe5 41.fxe5 f4 42.Qd4 with a draw.

Also possible was 35...Rc5 36.Rg3+ Bg7 37.h3 h5 38.Kh2 Kh7 39.Qd2 Bh6 40.Rgc3 Qe8 41.Qf2. At the end of this variation Fritz finds the position almost equal, just slightly in favor of Black (-.27) (16 ply).

In the game continuation, after 35...Rc5 36.Rg3+ Bg7 37.h3, Morphy now goes wrong, playing 37...Kh8?

As pointed out by Sergeant, Anderssen now had the opportunity of playing 38.Qd2! Instead, Anderssen played 38.Rxg7?.

Sergeant remarked, <A sacrifice for which there was no necessity, since Qd2 left him with the superior game still in every variation.>

Sergeant is correct. After 37...Kh8 38.Qd2, White has a strong advantage after every plausible defense by Black: 38...Rc4; 38...Qe8; 38...Qc6 or 38...Qd7.

Anderssen has just missed his last winning opportunity for this game.

After 37...Kh8 38.Rxg7, it is an equal position (.00) (19 ply) 38...Rxg7 39.Rc3 e3 40.Qf6 Rxc4 41.Qf8+ and the game would be a draw.

Later Steinitz erroneously condemned Morphy's move 39...e3. Actually, 39...e3 was the only move that draws for Black, but Steinitz claimed a win for Black with 39...Rc8. However, after 39...Rc8, White has a clear win by playing 40.Qf6 Qe8 41.d7 Qxd7 42.Ne5 Rxc3 43.Nxd7.

40.Rxe3 was a tragic blunder. 40.Qf6 was a clear draw. Anderssen may have missed that his f4 pawn could be taken with check on move 42.

A most interesting game, despite all the missed opportunities.

37...Kh8?; 38.Qd2! Qc6 (1.72) (18 ply) 39.Rd1 Qd7 40.Ne5 Rxc5 41.Rxg7 Re8 42.Rxf7 Qxf7 43.Qc3+. Now the evaluation is at (2.70) (18 ply) and the suggested continuation is 43...Qg7 44.Qc4 Qd7 45.Qd4+ Qg7 46.Qd5 Rg8 47.g4.

37...Kh8?; 38.Qd2! Qe8 (2.26) (18 ply) 39.Ne5 Bxe5 40.fxe5 Qe6 41.Rxc5 bxc5 42.Qb2 Rf8 43.Rb3 a6 44.Rb7 e3 45.Re7 Qd5. Or if 40...f4 41.Rxc5 bxc5 42.e6 fxg3 43.exf7 Qxf7 44.Qc3+ Kg8 45.Qxg3+ Kf8 46.Qe5 Qd7 47.Qh8+ Kf7 48.Qxh7+

37...Kh8?; 38.Qd2! Rxc4 (2.38) (18 ply) 39.Rxc4 Qxc4 40.d7 Rf8 41.d8(Q) Rxd8 42.Qxd8+ Qxg8 43.Qd7 Qf8 44.Qxa7 Qf6.

37...Kh8?; 38.Qd2! Qd7 (2.52) (19 ply) 39.Ne5 Bxe5 40.fxe5 Qe6 41.Rxc5 bxc5 42.Qb2 Rf8 43.Rb3.

Anderssen 5 mistakes:
29.Qb2+ 0.37 (29.Rh3 2.31)
35.Rc1 -0.39 (35.Rg3+ 0.00)
38.Rxg7 0.00 (38.Qd2 1.26)
40.Rxe3 -5.59 (40.Qf6 0.00)
41.Qf6 #4 (41.Qe5 -5.75)

Lasker also writes, "Anderssen, who in seeking combinations merely follows his instincts, does not guide himself by the form and nature of weaknesses and therefore does not discover the strongest line of play.

Obviously, the h-file is a Black weakness, also, the second and third [Black] ranks.

On the other hand, the position of the White knight, queen, and f1 rook on the same diagonal would invite Qb5, if only the rook on b3 were away, for the knight on c4 is insecurely posted, the d5 pawn is weak.

Therefore, not 28. Rh3 as indicated by a well-known book on Morphy, but first evaluation of the knight on c4."

Thanks for the link. I'm sure Sutovsky is very knowledgeable among his peers but (for those of us who don't have GM-level careers to sustain while also serving as FIDE DG) he makes some interesting mistakes.

Probably the most notable is that he conflates Daniel Harrwitz and Bernhard Horwitz -- thinks they are one guy. Harrwitz was probably the better OTB player (Harrwitz - Horwitz (1846)) but Horwitz did some amazing endgame studies -- hate to see him forgotten or, I guess, merged.

In Sutovsky defense, Sagar just asked him to start anywhere in the whole history of chess, and he decided on air that it should start with Morphy. Obviously he should have prepped.

Hopefully, he will correct some of those when he gives another talk. The Horwitz-Harrwitz mistake is a common one, but glaring nonetheless.

He also left out Morphy's whole experience in England, and the Loewenthal match, which he could have talked about during his digresssion on Staunton.

But his enthusiasm about this game 6 is wholly justified. Morphy seems affronted by 1 a3 and goes on the offensive almost from the start culminating in the unwise 11...g5

He played it here in the 10th game with his king on h8 and Anderssen's bishop on the long diagonal:

click for larger view

Anderssen vs Morphy, 1858

Ivanchuk gives a lengthy and fascinating review of this game at https://m.youtube.com/watch?v=MKy5i... Starts at around the 32 minute mark, has some minor audio issues that seems to have discouraged viewers, but is a fresh and non-computer look.

He is attracted by 28 Be5 as was Lasker, and like Lasker he considers 28 Rh3 only to give up on it when he sees that it gives up the b5 square to the queen. So we may deduce that the computer 28 Rh3 is very hard to calculate.

"Anderssen, who in seeking combinations merely follows his instincts, does not guide himself by the form and nature of weaknesses and therefore does not discover the strongest line of play.

This appraisal rings true about moves 28-31. Anderssen chooses a top tier move to start 28 Bxf6+ although not the absolute top move 28 Rh3.

29 Qb2+ is still a very good move, but significantly worse than 29 Rh3. It still wins, but the path to a win gets narrower.

30 Rh3 is the best move for the third move in a row, but now after 30...Rg7 White has only one move to maintain the win

31 Qd4 throws away the win, although it looks totally natural, and even Ivanchuk reviewing the game did not question it.

Anderssen would have to switch gears, and find 31 Qb3,

click for larger view

and win a long technical endgame after the non-obvious sequence-