TCEC Season 17 CPU League 1 Game 302: Laser-rofChade "Laser's Immortal"

Game	Season 17, CPU League 1, Game 302
Result	1-0
ECO	A37
Links	TCEC archive Lichess

This game was named "Laser's Immortal" by ICCF Grandmaster and the 28th ICCF World Champion Leonardo Ljubičić.

Game overview

                                                                                          ♛ ♜♚       ♟♝  ♟  ♜♝    ♞♟ ♟♘ ♟♕    ♙       ♘   ♙  ♙♙   ♙♗     ♖♖ ♔

Position after 23... Rd6

1. c4 c5 2. g3 g6 3. Nf3 Bg7 4. Bg2 Nc6 5. O-O d6 6. Nc3 e6 7. e3 Nge7 8. d4 O-O

Book exit into a well-known position. This position has been seen some dozen odd times at the super-GM level.

9. Re1 cxd4

The last move is interesting. Typically, the center tension is kept for a bit longer, and moves such as 9... Rb8 or 9... Re8 would be played, instead.

10. exd4 d5 11. cxd5 exd5 12. Bg5 h6 13. Bxe7 Nxe7 14. Qb3(N) Rb8 15. h4 a6 16. Ne5 Be6 17. Ne2 g5 18. hxg5 hxg5 19. Rad1 Nc6 20. Nc3 Na5

Black prepares to get the knight on c4 with tempo, but delays the move a bit, until the knight has to move.

21. Qc2 b5 22. Qe2 Rb6 23. Qh5 Rd6 (diagram) 24. b4

In such positions, the question is always about the long-term plan. 24. b4 is a somewhat committal move, although not necessarily a bad one. White could have also played 24. b3, controlling the c4 square and questioning the knight's positioning.

An interesting option might have been to play a useful waiting move provoking a line such as: 24. Rc1 Nc4 25. b3 Nxe5 26. dxe5. This would changed the game towards more double-edged direction, increasing the attacking potential for white at the cost of passed central pawn for black.

                                                                                           ♝♛ ♜♚       ♟♝  ♟  ♜       ♘♟♘ ♟   ♙♞♙        ♕ ♗♙      ♖♙       ♖ ♔

Position after 34. Rce2

24... Nc4 25. a4

White pushes the a4-pawn to undermine the protection of the knight on c4. Black is not really keen on letting white to take on b5, since 25... Rb6 26. axb5 axb5 and the white knight on c3 is pressuring both the b5 and d5 and pawns. Further, a continuation such as 27. Nf3 f6 28. Nh2 Bf7 29. Qf3 would offer white some additional initiative along with the open a-file, as black needs to put some effort on defending the b/d pawns and the kingside.

Therefore, black exchanges a4-pawn straight away.

25... bxa4 26. Nxa4 Bf5 27. Nc5

This is multi-purpose move, threatening to fork on b7 winning a tempo, also closing access to the c-file for black, and coordinating with the knight on e5. The two white knights are now annoying black with the light-square control and fork potential. It should also be noted that the f7-square is pressured by white, limiting the options for the black rook on f8.

27... Bc8 28. Qf3 Nb2

This move has to be a waste of tempo, since where is the knight going? The poke at the rook on c1 simply asks the rook to reposition itself, which it was probably going to do anyways sooner than later. Often such moves are a sign of engines not having a clear plan.

White plays the rook to d2, which is a flexible move winning a tempo on the knight.

29. Rd2 Nc4 30. Rc2

Also other ideas were possible, such as doubling the rooks on the e-file. The rook on c2 threatens to win a pawn with Nxc4 later, and black creates a counter-threat to win the pawn back.

30... Qb6 31. Qb3

Taking the pawn here would be premature: 31. Nxc4 dxc4 32. Rxc4 Rxd4 and black won the pawn back.

31... Qd8 32. Qd3

White is adding protection to the d4-pawn against the counter-capture idea.

32... Bf6 33. Bf3 Bg7 34. Rce2 (diagram)

Since adding the pressure directly does not seem to work, white shifts the focus on the e-file. The threat now is Re7 and Bh5 to pressure the f7 pawn, and to use that leverage to double the rooks on the 7th rank.

To illustrate the threat, consider the following line: 34... Rf6 35. Bh5 Qc7 36. Ng4 Rf5 37. Re7 Qd8 38. Ra7 Qb6 39. Ree7. Since this is clearly out of the question for black, black would have to give up the center control with 35... Nxe5 36. dxe5 which does not look promising, either. Black tries to defend against this threat.

                                                                                            ♝♛ ♜♚       ♟♝  ♟      ♜   ♘♘♖♞    ♙♕♙  ♟        ♙       ♙♗      ♖ ♔

Position after 42. Re5

34... Rh6 35. Bg2 g4

The pawn on g4 puts a hard stop for the bishop entering the h5 square, but it is a committal move.

36. Qc3 Bf5 37. Qc1

As the queen sidesteps the x-ray of the Bg7 bishop, the Nxd4 threat winning a pawn is now renewed. To illustrate why this was necessary, consider the following variation:

• 37. Nxc4 dxc4 38. Rd1 Rd6 39. Nb7?? Rxd4 and black is winning because the queen on c3 is pinned.

However, now as the queen is on c1 protecting the rook on d1, the rook cannot enter the d6 square because of the Nb7 fork. Also, the direct Bxd4 counter-capture would not work either. Consider the following line for the rationale: 37... Kh7 (wasting a tempo) 38. Nxc4 dxc4 39. Rd1 Bxd4. The black queen is now pinned behind the bishop, and white can take full advantage of the exposed black king. 40. Ne4 Bxe4 41. Bxe4+ f5 42. Qf4 fxe4 43. Qxe4+ Kh8 44. Rxd4. In this variation, white will be pawn or two up, easily winning against an open king.

Therefore, this black bishop retreats to protect the b7 forking square.

37... Bc8 38. Ned3

Now the rooks threaten to enter the 7th again, so the knight is repositioned to f5 via d6 to control the e7 square.

38... Nd6 39. Nf4

White concretely threatens the d5 pawn now and there's not much that black can do about it.

39... Nf5 40. Nxd5 Rd6

Black cannot win the pawn back, since white's initiative would be too strong: 40... Nxd4 41. Ne7+ winning an important tempo in clearing the e-file for the pin. 41... Kh8 42. Rd2 (pinning the queen) Qb6 43. Re4 Rd6 44. Rexd4 Bxd4 45. Ne4 Re6 46. Nxc8 Qxb4 47. Ne7 Qxe7 48. Rxd4 with two minors for a rook for white, and open king for black. Unfortunately, black did not have 43... Qxb4?? 44. Rexd4 Bxd4 45. Rxd4 leaving the black rook on h6 hanging: 45... Qxd4?? 46. Qxh6#

41. Qc4 Rh6 42. Re5 (diagram)

White offers the exchange, which we will see shortly. Before that, the black king sidesteps the Ne7+ check, the threat being 43. Rxf5 Bxf5 44. Ne7+ Kh8 45. Nxf5.

                                                                                                         ♚        ♜ ♟    ♘     ♘       ♙ ♙  ♟        ♙       ♙       ♜♗♔

Position after 50. Nxf6

42... Kh8 43. Qd3

Now white threatens to win two pieces for a rook with Rxf5, which prompts black to accept the exchange sacrifice. The main alternatives were:

• 43... Nd6 44. Nxa6
• 43... Qg5 44. Re8 Nd6 45. Rxf8+ Bxf8 46. Nxa6

Both variations losing another pawn for very little compensation.

43... Bxe5 44. Rxe5

Now the simple threat is to win two pieces for a rook by taking the knight on f5. The main alternative would not have been any better for black: 44... Nd6 45. Ne7 Re8 46. Qe3 Kg7 47. Bd5. Now black needs to give back the exchange in a difficult position with 47... Rxe7 to avoid white crushing through. For example: 47... a5 48. Nc6 Qf6 49. Rxe8 Nxe8 50. Qxe8 would leave white with decisive material advantage.

44... Qd6 45. Rxf5 Bxf5 46. Qxf5 Re8

Black gives up the pawn on f7, since trying to hang on to the pawns would not have worked anyways. For example: 46... Qg6 47. Qe5+ Qg7 48. Qf4 Qg6 49. Ne3 Qb1+ 50. Nf1 Qg6 51. Nd7 Re8 52. Ne5 and the knight and the queen are hitting both the f7 and g4-pawns.

47. Qxf7 Re1+ 48. Bf1 Rh7 49. Qf6+

White forces the queens off the board to enter the bishop+knight+knight vs rook pair endgame.

49... Qxf6 50. Nxf6 (diagram)

Both remaining black pawns are doomed. The pawn on a6 is attacked twice and its forward square is controlled by the pawn on b4. Similarly, the g4-pawn is hanging, and even if black defends it, a simple plan of Be2 will add more pressure after the king sidesteps the pin, of course. Black gives up the pawns to win one back.

                                                                                                          ♜ ♚                    ♙♘        ♗   ♙      ♘     ♜   ♙         ♔

Position after 59. Bc4

50... Rg7 51. Nd5 Rf7 52. Ne3 Rb1 53. Nxa6 Rb2 54. Nxg4 Rd7 55. Nc5 Rd6

Black could have taken one pawn back already here, but decided to wait one more move.

56. Nd3 Rxd4

Now that the position is clarified, it is worth considering the plans for both sides. White obviously wants queening a pawn, or win a rook to prevent queening. But this is easier said than done, since the knights are clumsy pieces to escort a pawns and maintain coordination against the rook pair.

Black has two main paths to a draw:

1. Create threats with coordinated rooks to win two minor pieces for a rook. Such threats happen naturally when the pieces are protecting each other. Since there are not many pawns left on board, there are also not that many protected squares. This is particularly problematic for the knights, since the two remaining pawn islands are separated enough so that a knight on a protected square on one pawn island cannot cover pawns on the other island.
2. Win a pawn. A coarse rule of thumb is that a rook can be sacrificed for a minor and a pawn in an endgame, especially when the pawns are not very advanced. Thus, if only two pawns remain, then two such exchange sacrifices would be enough to secure a draw. However, this is not quite that straightforward here, since it depends what pawn is won. If the b-pawn is lost, the connected f/g-passers might still be enough for white to win.

By the coarse rule of thumb, white would be winning (and computer analysis confirms this.) White begins to push the pawns on both fronts.

57. b5 Rd8 58. g4 Rf8 59. Bc4 (diagram) Rd8

Here black could have won the f2-pawn. Not of course with Rfxf2?? since 60. Nd3 would fork the rooks.

However, taking with the other rook was more interesting: 59... Rbxf2 60. Nd7 (only move, hitting rook on f8) R8f4 61. Ne5 (another only move, threatening fork on d3) 61... Rf6 62. g5 Rf8 63. Ng6+ forking. This pawn grab would have ultimately failed, since white would have been able to take away all the squares on the f-file for the other black rook, before black had a chance to move the rook on f2 into safety.

                                                                                                               ♗              ♚      ♙                    ♘♔        ♙

Position after 74. Bxg8

60. g5 Rb4 61. Bf7 Kg7 62. g6 (anchoring the bishop) Rb1+ 63. Kg2 Rc8 64. Ne6+ Kf6 65. Nd5+ Ke5 66. g7 Ra1

66... Kxd5 would have been no good due to 67. Nf8+ exposure check with queening on the next move.

67. Nf8

This is a typical idea to protect the queening square by intercepting the rook's visibility.

67... Rcc1

Desperate attempt to prolong the game. If now 68. g8=Q, then 68... Rg1+ would be able to exchange one rook for the queen. But white has other plans.

68. Kf3 Ra3+ 69. Ne3 Rg1 70. Ng6+

Another interception, but this time it is with a check. Now black must give up an exchange to prolong a queen appearing on the board.

70... Rxg6 71. Bxg6 Ra8 72. Bf7 Kd6 73. g8=Q Rxg8 74. Bxg8 1-0. (diagram)

White wins by the 6-men endgame tablebase adjudication.

Bishop and knight and two pawns against a lone king is trivial win. With best play on both sides, this will be mate after 11 more moves. For example: 74.... Kc5 75. Bc4 Kb6 76. Nd5+ Kb7 77. b6 Kc6 78. Ke4 Kb7 79. f4 Kc6 80. f5 Kb7 81. Bb5 Kb8 82. f6 Kb7 83. f7 Kb8 84. Ba6 Ka8 85. f8=Q#.