My second knight ending from our club online tournament a couple of weeks ago was, in truth, one of the most boring games I’ve played in the past 60 years, and, as most of my games are extremely boring, that’s quite an achievement.
I guess you must be reasonably happy if you get a Weighted Error Value: White=0.16 (very precise) /Black=0.13 (very precise) after a 45 move speed game, but in fact nothing happened.
I don’t know who my opponent was: I think a new member. Playing White, he traded everything off at the first opportunity, admittedly with a lot of help from me, reaching a knight ending, which, unlike the game I showed you last week, was, frankly, rather dull.
There were two points of interest, deriving from what might have happened rather than what actually happened.
Here’s the position at the start of the ending, where I have to play my 25th move.
25… Kc6
26. Nd3 Nd7
27. f4 b5
28. Kf2 bxa4
29. bxa4
We have a familiar pawn barrier on the queenside where neither king can enter, so I tried to break through in the centre.
29.. f6
30. g4 Kd6
31. h4 e5
Taking once on e5 is fine, but if my opponent trades everything on e5 he’s losing the pawn ending. For example 32. fxe5+ fxe5 33. dxe5+ Nxe5 34. Nxe5 Kxe5 35. Kf3 h6! (the only winning move) 36. g5 hxg5! 37. hxg5 Kf5!. My opponent’s choice, f5, is also absolutely fine.
32. f5 gxf5
33. gxf5 e4
34. Nb2 Nb6
35. Kg3 h5
36. Kf2 Nc8
37. Kg3 Kc6
38. Kf4 Nd6
39. Nd1 Nc4
40. Nc3 Nb2
41. Kg3 Nc4
42. Kf4 Nb2
43. Kg3 Nc4
44. Kf4 Nb2
45. Kg3
With a draw by repetition.
1/2-1/2
But the really interesting question is this. What would have happened if I’d played Nxa4 rather than Nc4 on move 41 or move 43.
Here’s the position after 43.. Nxa4!?.
How would you assess this? Will White’s extra knight win, or will Black’s a-pawn promote? We all know that knights struggle against rooks’ pawns.
Place your bets now.
44. Nxa4 Kb5
45. Nc3+!
The check gains an important tempo: Nc5 is too slow: 45. Nc5 a4 46. Nd7 a3 47. Nxf6 a2 48. Nxd5 a1=Q 49. Kf4 Qe1 50. Kxe4 Qxh4+ 51. Nf4 Qe7+ 52. Kf3 Qg5 and wins.
45… Kc4
46. Na4 Kb4
47. Nb6 a4
48. Nxd5+!
Not 48. Nxa4? Kxa4 when Black wins the pawn ending: 49. Kf2 Kb3 50. Ke2 Kc3 51. Kf1 Kd3 52. Kf2 Kd2
48.. Kb3
49. Nxf6 a3
50. Nxe4 a2
51. f6 a1=Q
A fascinating position. White’s captured three pawns while Black has promoted.
52. Kf4!
This and Ng5 are the only drawing moves. Another sample variation: 52. Ng5 Qf1 53. f7 Qf5 54. Kg2 Kc4 55. e4 Qf4 56. e5 Qg4+ 57. Kf1 Qf4+ 58. Kg2 Kd3 59. e6 Ke2 60. Nh7 Qg4+ 61. Kh2 Qxh4+ 62. Kg2 with a perpetual check.
52.. Qh1
53. Kf5 Qxh4
54. Ng5!
The only drawing move.
54.. Qg4+
55. Kg6 h4
56. f7 Qc8
57. d5 Qb8
58. e4 Kc4
59. Ne6 h3
60. f8=Q Qxf8
61. Nxf8 h2
62. d6! h1=Q
Black’s third queen, but it’s still not enough to win.
63. d7! Qxe4+
64. Kf7 Qd5+
65. Ke7 and it’s a draw. Black has no way of making progress.
If I’d played Nxa4!? a very boring ending would have been transformed into something really exciting. But it hadn’t occurred to me at the time, and, even if I’d looked at it I’d probably have chickened out anyway.
Would you have taken the risk?
What can you learn from this? I guess that knights are unable to stop passed rooks pawns supported by a king. Also that sometimes an ending that looks drawn has unexpected winning chances, and that sometimes an ending that looks lost might possibly have a miracle save.
In the immortal words of Bob Dylan, ‘Ain’t it just like the knight to play tricks’.
Richard James
