Missed opportunities

Gordon W. Scott - Peter J. Batchelor [E04] British Championship 2015 1.d4 Nf6 2.c4 e6 3.g3 d5 4.Nf3 dxc4 5.Bg2 Bb4+ 6.Bd2 c5 7.Bxb4 cxb4 8.Ne5 0-0 9.Nxc4 Nc6 10.e3 e5 11.d5 b5 12.0-0 bxc4 13.dxc6 Qb6 14.Nd2 Ba6 15.Qc2 c3 16.bxc3 Bxf1 17.Rxf1 a5 18.cxb4 axb4 19.Rc1 Qc7 20.Qb2 Rfb8 21.Nc4 Rb5 22.Rd1 e4 23.Nd6 Rd8 24.Qc2 Rb6 25.Nxe4 Rxd1+ 26.Qxd1 Rxc6 27.Nxf6+ Rxf6 28.Qa4 Qc5 29.Qe8+ Qf8 30.Qxf8+ Kxf8 31.Bd5 Rf5 32.Bb3 Rc5 33.f4 Ke7 34.h4 f6 35.Kg2 Rc1 36.Kf3 h6 37.Kg4 Rf1 38.Bc4 Rf2 39.h5 Rc2 40.Bd5 Rc3 41.Kf3 Rc5 42.Ke4 Kd6 43.Bf7 f5+ 44.Kf3 Rc3 45.Bb3 Kc5 46.Kf2 Kb5 47.Ke2 Rc1 48.Kf2 Rh1 49.Kg2 Rb1 50.Kf3 Kc5 51.Bf7 Rb2 52.Bb3 Rh2 53.Bf7 Kb5 54.Bb3 Ka5 55.Bf7 Ka4 56.e4 fxe4+ 57.Kxe4 Rg2 58.Kf3 Rg1 59.g4 Ka3 60.Be6 Rf1+ 61.Kg3 Kb2 62.g5 Kc3 63.gxh6 gxh6 64.Kg4 Kd4 65.Kf5 Ke3 66.Kg6 Kxf4 67.Kxh6 67...Ke5 Best; the only other winning move is 67...Rg1 (one move slower). 68.Bb3 As good as anything. 68...Kf6 Again best, the alternative being ...Rg1. 69.Bc4 It looks natural to defend a2 and attack the R although the tablebase says Ba4 or Bc2 last one move longer. 69...Rf5 Many moves win but Rd1 is best. 70.Bg8 Relatively best (equal with Bb3). 70...Ra5 One of several moves that mate in 37. 71.Bc4 Ra7 [71...Ra8 (threatening mate on h8) is the right idea although it is in fact only one move quicker says Probe. The point is that White must then play Kh7 as eventually happens in the game.] 72.Bd5 White can only wait and hope Black cannot find the winning plan. 72...Rc7 [72...Rd7 73.Bc4 Rd8 (again threatening mate) is one move quicker.] 73.Bb3 Rc8! At last finding the right idea to move on to the next phase of the solution. 74.Kh7 Rd8! Before playing ...Kg5 (which wins more slowly) Black wants his rook on a dark square. [74...Kg5? 75.Kg7! and White can draw, e.g., 75...Kxh5 The black king has been dragged too far from the queenside, it seems. 76.Be6 It was to avoid this tempo-gaining move that Black first played ...Rd8 before ...Kg5. On e6 the bishop creates a temporary barrier that lengthens the Black king's path to the queenside: a subtle but crucial difference from the main line. 76...Rc5 77.Kf6 Kh4 78.Bb3 Kg4 79.Ke6 Kf4 80.Kd6 Rc1 (80...Rb5 81.Bc4 Rh5 82.Bd5!? Several moves draw in fact, now the White king is near enough to the pawns. (A simpler way is 82.Kc6 Ke3 83.Bb3 Kd2 84.Kb6 Kc3 85.Ka6 By no means the only safe move 85...Kb2 86.Kb6 Ka3 87.Ka6 White draws as long as only one file separates his king from the Black pawn. 87...Rh2 88.Kb5 or Ka5 88...Rh3 (88...Rb2!? should also be met by a B move such as Be6 but not 89.Kc4?? Rb1 zugzwang.) 89.Bc4= or Be6 or Bf7 or Bg8. ) 82...Ke3 83.Kc5 Kd3 It might look as if the self-pin has got White into trouble. 84.Kc6! The unpin saves White. 84...Kc3 85.Bb3 Kb2 Black's plan is to win the king and pawn ending, but... 86.Kb6! Rh3 87.Bc4 The B runs away and Black cannot make progress. Note however that Be6?? loses to Rh6 and also Bd5 loses: (87.Bd5?? Rh6+! forcing the king to the seventh rank, and after 88.Kb7 Ka3 89.Bg8 Rh2 90.Kb6 Rxa2 etc.) ) 81.Kd5 Ke3 82.Bc4 Also Ba4 holds. 82...Kd2 83.Kc5 Kc3 84.Bb3 and Black cannot win because his own pawn is vulnerable.] 75.Bc4 Rb8 This does not spoil the win but Black could have played ...Kg5 at once. [75...Kg5 76.Kg7 (76.Bf7 as in the game is slightly better but still loses.) 76...Kxh5 77.Kf6 Kg4 78.Ke5 (78.Bb3 Re8 is the start of the tablebase's longest variation.) 78...Kf3?! and the White king cannot get near the Black pawn in time.] 76.Bb3 Rf8 76...Rd8 is one move quicker says the tablebase. Black's idea with the text was presumably to deny Bf7 in reply to ...Kg5. 77.Bg8!? Ingenious though it makes no difference to the result according to the tablebase. 77...Kg5?? Without this Black cannot make progress but he should have the R on d8 first. Now the game should be drawn because White plays Kg7 with tempo. [77...Rd8! returns to the winning path: 78.Bc4 Kg5 79.Bf7 Rd2 80.Kg7 Rh2 Zugzwang! 81.Kf8 Trying to get over to the queenside. (81.Kg8 loses one move faster.) 81...Kf6 82.Ke8 Re2+ 83.Kf8 Re5 84.h6 The h-pawn cannot be saved, e.g. (84.Be8 Re7 85.Bc6 Rc7 86.Be8 Ra7 87.h6 Rh7 88.Kg8 Rxh6 ; 84.Bg6? Ra5 ; 84.Kg8? Ra5 ) 84...Re7! 85.Bb3 Rh7 86.Ke8 Rxh6 87.Kd8 Ke5 88.Kd7 Kd4 and compared with the drawing lines earlier, the White king cannot get close enough to the b-pawn.] 78.Kg7! Rf6 79.Bf7 Ra6 80.Kh7?? This should have been the losing move and the first move on the winning path is not difficult to see. Black had about 7-8 minutes left (plus increments) and thought for most of his time before deciding he could not win and played Rxa2. However... [80.Bc4[] White establishes a fortress defending his pawns and may even advance h5-h6 if Black is careless. 80...Ra7+ (80...Rh6 81.Bf7[] ; 80...Rd6 81.Bf7 Rc6 82.Bg6 ) 81.Bf7 ] 80...Kf6! Black mates in 36. [Instead of 80...Rxa2? 81.Bxa2 Kxh5 82.Kg7 b3 83.Bxb3 with a draw, as actually occurred.] 81.Bb3 81 Bg8 loses in the same number of moves and others are worse. [81.Bg8 Rd6 82.Kh6 Rd8 83.Kh7 Kg5 84.Bf7 Rd2 85.Kg7 Rh2 winning as in the 77...Rd8 variation] 81...Rd6 82.Bc4 [82.Kh6 Rd8 etc.] 82...Kg5 83.Bf7 Rd2 84.Kg7 Rh2 This L-shaped rook manoeuvre is again the key to winning the position. The double attack on the a-pawn and h-pawn force the White K to abandon its g7 square after which zugzwangs and double threats eventually force the h-pawn to advance to h6 where it is inevitably lost. Then the white king on the edge cannot get to the drawing zone a6-b6-c6 where it needs to be when Black tries to win by an exchange sacrifice on a2 or b3. 0-1