"...a brilliant combination bewitches men to such an extent that they willingly believe falshoods and are blinded to the truth."
Emanuel Lasker, Lasker's Manual of Chess (1932)
In the Evergreen Game, Anderssen -- Dufresne, Berlin 1852, this position was reached after White's 19.Rad1!?*
Black to move
Dufrense played 19...Qxf3, which is obviously a mistake. Of course, it is obvious to us because we know Anderssen's winning combination. Dufresne threatened checkmate in one. Did he play this move out of despair because he fully understood the implications of Anderssen's last move?
Black has alternatives that require study if I am to meet Rashid Ziyatdinov's training standard (see "To Know a Position"). Diagram 146 in GM-RAM is this game before 19.Rad1. These alternatives may be explored in several ways: playing the position against a friend, carbon or silicon; watching chess videos (there are several good ones); or reading books that analyze this historic game.
19...Qh3 also threatens checkmate in one, but has the additional benefit of defending the pawn on d7.
19...Rg4 threatens the queen, frees g8 for the king's escape, and prepares interference plans on the fourth rank.
Practice against the Computer
I tried 19...Qh3 this morning against Hiarcs on my iPad.
That game continued 20.Bf1 Qf5 21.c4 Rg4
21...Rg6 was Stockfish's most significant improvement to my play, with rough equality, according to my silicon friend.
22.Qb4 Rf4 23.Bxe7 Qxb5 24.cxb5 Nxe7 25.Rxe7 Kf8 26.Rdxd7 Rxf6 27.Bc4
Black to move
I looked at 27...Rd8, but failed to calculate as I knew that I was already lost.
28.Rxf7 Rxf7 29.Rxf7 Ke8 30.Rxh7 Rd8 31.Rh8+ Ke7 32.Rxd8 Kxd8 33.gxf3 1-0
19...Rg4! was suggested by Lasker in Lasker's Manual of Chess. He notes, it "causes difficulties" (272).
Graham Burgess offers the view that 19...Rg4 "is the best try" (The World's Greatest Chess Games , 23). He develops several lines in his analysis.
a) 20.c4 Rxg2 21.Kxg2 Qg4+ 22.Kf1 Qxf3
White to move
Then the combination as in the game: 23.Rxe7+ Nxe7 24.Qxd7+ Kxd7 25.Bf5+ Ke8 26.Bd7+ Kf8 27.Bxe7+ "is no longer mate, because Black has the g8 square at his disposal."
Alternately, 23.c5 Qh3+ 24.Kg1 Ne5 "and it is Black who is attacking."
b) 20.Re4 "the key line". 20...Rxe4 21.Qxe4 "although White's threats aren't too devastating, ... it is difficult for Black to find a decent move." Burgess continues with analysis of several possibilities, all favoring White.
Lasker offers three brief lines after 19...Rg4.
a) The combination as in the game: 20.Rxe7+ Nxe7 21.Qxd7+ Kxd7 22.Bf5+ Ke8 and his variation ends here, but he notes that g8 is available to the king, so checkmate is no longer a threat.
b) 20.Qc2 (Lasker mentions the queen moving, but does not offer the square) 20...Rxg2. His line ends here, but it is easy to imagine that he considered the moves suggested by Burgess after 20.c4.
c) 20.Be4 d5.
Because the soundness of 19.Rad1 is put into doubt, Lasker concludes, "the aesthetic value of Anderssen's move has as yet not been demonstrated" (272).
Garry Kasparov, My Great Predecessors. Part I (2003) mentions Lasker's analysis, which followed efforts by Paul Lipke, Deutsche Schachzeitung (1898). He also mentions the efforts of Hoppe and Heckner (no citation) who, "tried to demonstrate a win for White" in 1930 (28). Kasparov presents two lines from this work. The first begins with 20.c4 and follows the moves presented by Burgess.
The main variation, according to Kasparov, begins with 20.Bc4.
20...Qf5 21.Rxd7 Kxd7 22.Ne5+ Kc8 23.Nxg4 Nd5 24.Qd1
Black to move
Kasparov indicates that Hoppe and Heckner continue 24...Nd8 25.Bd3 Qd7 26.Ne5 Qe6 27.Nxf7++-.
He offers an improvement:
24...Nxf6 25.Bd3 Qxg4 26.Qxg4+ Nxg4 27.Bf5+ Kd8 28.Rd1+ Nd4 29.Bxg4 Bd5 30.cxd4 Bxa2 "with a roughly equal endgame," according to Kasparov.
The Way Forward
There is much to do in my study of the Evergreen Game.
Memorizing the game itself and Anderssen's final combination is easy enough. I already recognize the position before 19.Rad1 as a fingerprint of the game.
Burgess and Kasparov's annotations offer insights into the opening, middlegame, and concluding combination that merit study and practice.
There are lines after 19.Be4 that I have yet to examine.
Today's work on the possibilities after both 19...Rg4 and 19...Qh3 only scratch the surface. It should be clear that at a pace of one game per week, there will be plenty of work remaining on this game after I have moved on to the next.
*Emanuel Lasker points out, as have many writers in his wake, that 19.Be4 wins with fewer complications.