1.d4 Nf6 2.Nf3 e6 [White's coming set up may also be played if Black employs a kings-side fianchetto. Black, though, probably has an easier time of it as e5 is not availalbe for the White knight: 2...g6 3.Bg5 Bg7 4.Nbd2 0-0 5.e3 d6 6.c3 c5 is recommended, in contrast to the Nbd7 set up aiming for e5, by Avrukh 2012 (as well as Kornev 2016), who then claims Black soon gets at least easy equality after (Alternatively 6...h6 forces a decision about the bishop 7.Bh4 (7.Bxf6 is less flexible 7...Bxf6 8.Bd3 e5 9.Qb3 Kg7 10.0-0-0 Qe7 11.h3 c6 12.Ne4 Nd7 Donnelly-Naylor, Coventry League 1989 and now instead of 13. Nxf6?! simply 13.Rhe1 keeps the game balanced.) 7...c5 8.Bd3 cxd4 9.exd4 Nh5 10.0-0 Nf4 Speelman-Howell, London Classic Rapid 2016.) 7.Bxf6 Bxf6 8.dxc5 dxc5 9.Ne4 Qb6 etc.]
3.Bg5 Quite a typical opening choice for Tartakower who often played a very wide range of unusual or uncommon lines. Besides, Golombek was an extremely conservative player who favoured solid main line openings such as the Queens Indian and Nimzo-Indian against 1.d4 and played them repeatedly. Similarly, he mostly played the Caro-Kann against 1.e4 although he also ventured the Sicilian Dragon in the days when this was more of a positional opening and before the Yugoslav Attack become popular. In 1980 this White opening was also recommended by opening authority Leonard Barden.This was based on the Yugolsav player Klaric using it repeatedly to win the first prize of ÂQ1200 (from the tournament's main sponsor the National Bank of Dubai) in the 1979 London Evening Standard Congress. This would be worth around ÂQ6500 today.
3...Be7 4.Nbd2 d5 5.e3 c5 [Black may also omit this move and play instead 5...Nbd7 6.Bd3 b6 7.c3 0-0 8.b4 as occurred in Petrosian-Andersson, Amsterdam 1973.]
6.c3 Nbd7 [After 6...Nc6 an older game continued in a similar manner to the main game against a very formidable opponent but White got a strong attacking position after 7.Bd3 h6 8.Bh4 0-0 9.0-0 b6 10.Ne5 Nxe5 11.dxe5 Nd7 12.Bg3 Bh4 13.Bxh4 Qxh4 14.f4 Bb7 15.Rf3 as in Tartakower-Keres, Kemeri 1937.]
7.Bd3 b6 [A later Tartakower game went slightly differently 7...0-0 8.0-0 h6 9.Bh4 b6 10.Ne5 Nxe5 11.dxe5 Nd7 12.Bg3 f5 13.exf6 Bxf6 14.Qe2 Ne5 15.Ba6 Bxa6 16.Qxa6 c4 17.Rad1 Tartakower-Pirc, Bled 1950.]
8.0-0 Bb7 Black plays the opening in a flexible and subtle manner. He has developed solidly and delayed castling so that White's typical attacking set up is avoided. [After 8...0-0 now, or via being played earlier in the game, then White may continue 9.Ne5 Nxe5 10.dxe5 Nd7 11.Bxe7 Qxe7 12.f4 wth an pleasant game for White and which IM Bellin thinks is not easy for Black to equalise.]
9.Qc2 h6 10.Bh4 0-0 11.Ne5 White has played numerous moves in this position of which this and [11.Rfe1 are the most popular.; Whilst 11.Rae1 was not a problem for Black after 11...c4 in Spassky-Reshevsky, Amsterdam 1964.; However, 11.a4 for example as in Cast-Thiruchelvam, 4NCL 2001 expanding on the queens-side seems the most logical.]
11...Nxe5 [11...Rc8 is a little slow and about even after 12.f4 c4 13.Be2 Nxe5 14.fxe5 Ne4 15.Bxe7 Qxe7 16.Nxe4 dxe4 17.Bg4 Turpanov-Agnelo, Bled Olympiad 2002.]
12.dxe5 Nd7 Black has a fully acceptable game after this typical retreat following a capture on e5, but in this specific position [12...Ng4 preventing f4 is stronger with advantage to Black after 13.Bg3 Bh4 14.Bf4 (14.Bxh4 Qxh4-/+ Orhan-Nguyen Mink, GER-Tch Magdeburg 2010.) 14...g5 15.g3 gxf4 16.exf4 h5 17.gxh4 (17.Rae1 Be7-+ Wilder-Van der Strerren, Reyjkavic) 17...Nxh2 18.Kxh2 Qxh4+-+ Munar Rossello-Planas, Mallorca open 2000.]
13.Bg3 [13.Bxe7 Qxe7 14.f4 is level except after 14...f6 15.exf6 Nxf6+/= Halfmann-Brueggemann, Krefeld 2012.]
13...f6 [13...c4 with a general advance on the queens-side with b5 and a5 is more incisive and avoids a potentially weak e6 pawn.]
14.exf6 Bxf6 15.e4 Bh4 16.Rae1 Putting Black under a little bit of pressure along the e-file by subliminally attacking the weakness on e6
16...Bxg3 17.hxg3 c4 A trading of advantages. Black with tempo gains access to e5 and c5 for the knight whilst White has d4 for his knight.
18.Be2 Nc5 [18...Qc7 Connecting the rooks first seems a touch more accurate as the queen also controls c5 and e5 and supports c4.]
19.exd5 exd5 20.Nf3 Qf6 This, and Black's next, look very natural but with fine play Tartakower shows that Black can make little of the eventual pressure on f2. Meanwhilst White may dominate the key central squares d4 and e3. In addition, White can force Black to ceed the e-file due to the weak white squares of e8-h5. This is principally due to the somewhat passive position of the Bb7 which is blocked by, and reduced to defending, the pawn on d5.
21.Nd4 Rae8 [21...Rad8 holds more hope although Black's pieces are not very active.]
22.Bh5 Rxe1 23.Rxe1 Ne4 [23...Nd3 may be met by 24.Re2 when White threatens the annoying Qa4 as well as b3.]
24.Bf3 a6 25.Re2 b5 Generating the possibility of queens-side play with a5 and b4 but placing more pawns on white squares still means the Bb7 is passive.
26.Bxe4 dxe4 27.Qd2 Rd8 28.Qe3 White over-protects f2 and d4.
28...b4 29.Rd2 a5 30.Qe2 Qf7 31.Qg4 Controlling e6 and leaving Black the task of finding a decent plan.
31...Kh8 [The attempt to activate the bishop here with the pawn sacrifice 31...e3 fails to 32.fxe3 Rf8 33.Qe6+/- ]
32.Kh2 Rd5 33.Kg1 [White is forced back as obviously 33.Qxe4 loses immediately to 33...Rh5+ ]
33...Rd8 Possibly hoping for a repetition after 34.Kh2 Rd5 but White switches into an ending maintaining niggling pressure on Black.
34.Qf4 Qxf4 35.gxf4 Kg8 36.Kf1 A classic "grinding" position in which White can gradually improve the position of his pieces but Black is limited in what can be done.
36...Kf7 37.g3 Kf6 38.Ke2 g5 39.fxg5+ hxg5 40.Ke3 Rd6 In what appears to be a relatively quiet position Black has surpisingly few decent moves. If, for example, the natural king centralisation is chosen [40...Ke5 then 41.Nc6+ wins on the spot.; Or if 40...Rb8 trying to pentrate into White's position actually leads to the reverse. Some possible lines which show Black's probems are: 41.Rd1 bxc3 (and 41...Bd5 42.Nc6 Bxc6 43.Rd6+ Ke5 44.Rxc6 bxc3 45.bxc3 Rd8 46.Rxc4 Rd3+ 47.Ke2 a4 48.Rxa4 Rxc3 49.Ra5+ Kf6 50.a4+- ) 42.bxc3 Bd5 43.a4 Rb2 44.Rh1 Ra2 45.Rh6+ Ke5 46.Rg6 Rxa4 47.Rxg5+ Kd6 48.Nf5+ Ke5 49.Ne7+ Kf6 50.Rxd5 Kxe7 51.Rd4+- In addition to an extra pawn White will soon have 2 united pased pawns which are far stronger than Black's immobile remaining pawns.]
41.Nb3 cxb3 This sacrifice is actually fine for Black but is not followed up correctly. [Poor is 41...Rxd2 since 42.Nxd2 wins as White nets both the e4 and c4 pawns.; and 41...Rd3+ 42.Rxd3 exd3 43.Nd2 bxc3 44.bxc3 Bd5 45.Kd4 is problematic as the White king has free access to the a5 pawn.]
42.Rxd6+ Ke5? A shame. After a difficult game Black finally cracks. Advancing the king allows White to check, and gain a move to prevent the b3 pawn promoting, so White now wins. Instead [42...Ke7 just holds, for instance, 43.Rd1 bxa2 44.cxb4 axb4 45.Kd4 b3 46.Kc4 e3 47.fxe3 (47.Kxb3 Bd5+ snatches victory for Black.) 47...Be4 48.Ra1 (48.Kxb3 is a blunder too as Black wins after 48...Bb1 and the pawn promotes.) 48...Bc2 and White cannot win as Black can just play Bb1-c2 repeatedly. (but not 48...Bb1 49.Kxb3 Kd6 50.Rxa2 Bxa2+ 51.Kxa2+- ) ]
43.Rb6 bxa2 [43...bxc3 44.Rxb3 (Instead 44.Rxb7 loses to 44...bxa2 and the rook is overloaded.) 44...Bd5 45.Rxc3 Bxa2 46.Rc5+ winning readily.]
44.Rb5+ [Following 44.Rb5+ Ke6 45.Rxa5 catches the a-pawn in the nick of time and if Black guards it then 45...bxc3 (or 45...b3 46.c4 ) 46.bxc3 Bd5 47.Kd4 Bb3 48.c4 and in each case the White king gains access to c3 winning another pawn. (Or more simply 48.Kxe4 Bc2+ 49.Kd4 Bb1 50.c4 is also more than good enough. Chess can be a tough game but at least Golombek got revenge in 1949 and 1951 against Tartakower by which time he had obtained his IM title and also achieved two of his three British Championship wins.) ] 1-0