Karpova: Georg Marco's annotations condensed:
6...Nxe4 <It would be better for Black to concentrate on his development.>
7...c6 <The defense initiated thereby is not to Black's advantage, as the course of the game shows. Very much to be considered were 7...Nc5 or ...Ng5 to possibly meet Re1+ with ...Ne6. Bad would be 7...Nd6 or ...Nf6, e. g. 7...Nd6 8.Re1+ Be7 9.Bxg7 Rg8 10.Bf6 Nc6 11.Bd5 Kf8 12.Rxe7! (12.Bxc6 Bxf6!) 12...Nxe7 13.Nh4 Qe8 14.Qh5 Nef5 15.Nd2! and wins or 7...Nf6 8.Ng5 d5 9.Re1+ Be7 10.Ba3 Nc6 11.Bb5 Ng8 12.Qh5 g6 13.Qf3 Be6 (if 13...f6 or ...f5 then Nc3!) 14.Nxe6 fxe6 15.Bb2 and wins.>
8...d5 <With 7...c6 Black prepared covering the ♘e4, but is he really protected now? Couldn't White take him anyway as if he was en prise? Note 9.Rxe4+ dxe4? 10.Bxf7+ Ke7 11.Ba3+. Doesn't this prove that White could win the ♘e4? To answer yes would be too frivolous, as after 9.Rxe4+ Be7 or ...Be6 many complications arise whose chances for Black and White can only be assessed after long examination, e. g. 9...Be7 I) 10.Bxg7 dxe4 (or 10...Rg8!) 11.Qxd8+ Kxd8 (not 11...Bxd8 12.Ne5) 12.Bxh8 exf3 13.Bxf7 fxg2 with minimal material advantage for Black, barely compensating the lack in development; II) 10.Rd4 Bf6 who has the advantage? Only God knows that and perhaps Mr. Halumbirek.>
9...Bf5 <! If anything, the best. On all other moves follows 10.Rxe4! with decisive success, e. g.I) 9...Bd7 10.Rxe4! dxe4 11.Bxf7+ Ke7 12.Ba3+ Kf6 13.Qd4!
II) 9...Be6 10.Rxe4! dxe4 (if 10...dxc4 11.Nd2 with a strong attack) 11.Qxd8+ Kxd8 12.Bxe6 fxe6 13.Nf7+
III) 9...Nd7 10.Rxe4! dxe4 11.Bxf7+ Ke7 12.Ba3+ c5 (12...Kf6 13.Qd4!) 13.Qd5 Kf6 (If the Nd7 moves then Bxc5+, if 13...Qb6 14.Nxd7 Bxd7 15.Bxc5+) 14.Ng4+ Ke7 15.Qe6#
IV) 9...Be7 10.Rxe4! dxe4 11.Qh5! and wins.
V) 9...Bd6 10.Rxe4! dxe4 11.Qh5 g6 (if 11...0-0 then mate in 4) 12.Bxf7+ and wins.
VI) 9...Bc5 10.Rxe4 dxe4 11.Qxd8+ Kxd8 12.Nxf7+.
VII) 9...Qd6 10.Rxe4! dxe4 11.Qxd6 Bxd6 12.Nxf7 and wins.
But it should not be overlooked, that in variations I-VII only 10.Rxe4 dxe4 was examined. These few premises are not enough to damn Black's moves so far. Such a conclusion could only be drawn after exhaustive examination, e. g. in variation II the sideline 9...Be6 10.Rxe4 Qb6! as now three White officers are attacked.>
I corrected a typo in line V where the winning move is given as 12.<Lf7xg6> (Bxg6, which makes no sense), while 12.Bxf7+ was obviously intended and is crushing.
12.Rxe4 <! Finally, White can play his long-planned sacrifice.>
14...c5 <14...Kf6 is bad due to 15.Bxf8 Rxf8 16.Qd6+.>
19...Nc6 <! Important to prevent the future check on e5 and d4.>
23.Ba3 <!! Charming! Now Black sees many ghosts, as 23...Qb6
24.Rxc6!, if 23...Qg7 24.Qxh5! threatening Bb2!. Bad is also 23...Qd4 24.Qxd4+ Nxd4 25.Bb2 Bg7 26.Bxd4 Bxd4 27.Rc7 or 23...Qe5 24.Qxe5+ Nxe5 25.Bb2 Bg7 26.Bxe5 Bxe5 27.Rc5!. What abundance of horrors? Can Black be blamed for overlooking his guardian angel 23...Qe2! 24.Rxc6 bxc6 25.Qe5+ Bg7 26.Bb2!! (Black can neither take ♕ nor ♗) 26...Qg4+! 27.Kf1 Qd1+ 28.Kg2 Qg4+ with a draw.> The saving line may not be so good for Black as White needn't play 24.Rxc6 but can go for the ♕ exchange with 24.Qd1; in the analysis line, White further drifts into a bad position and Black could get an advantage after 27...Rg8 securing the ♗g7 (28.Bxg8?? Bxe5).
24.Rxc6 <!>
25.Rc7 <! Resigns. Nothing to be found against the threat Be6!.>
Source: Pages 33-34 of the January-February 1913 'Wiener Schachzeitung'