I showed you the first part of one of my online games, where we reached this rook ending after Black’s (my) 25th move.
All I have to do to win the game is play sensible moves.
I’ll skip through the next few without comment. They weren’t perfect on either side but didn’t change the assessment: Black is winning.
26. Re1+ Kd7
27. Rb1 Kc7
28. Re1 Kd6
29. Re8 Rxa2
30. Rd8+ Ke7
31. Rg8 Kf7
32. Rc8 Rxh2
33. Rxc6 a5
34. d5 g5
35. Rc7+ Ke8
36. d6 Kd8
Stockfish assesses this as about -9, but passed pawns are always dangerous. What was it Nimzowitsch said: “A passed pawn is a criminal who should be kept under lock and key”? One way of doing this would be to play Rh4 to stop the white king advancing. Another way would be g4 followed by Rh5. But instead I saw a tactical idea. If I forced the white king up the board I could advance my a-pawn safely because of the skewer.
This is exactly what I shouldn’t do, though. Once the king advances his pawn will give him promotion and mating threats. It should still be winning, but I’ve made my task a lot harder.
38. Kc4 a4?
This really does throw away the win. I should be using my rook from behind while pushing the g-pawn, not the a-pawn. 38.. g4 or 38.. Rh1, for example.
39. Kd5 a3
40. Ra8+ Kd7
41. Ra7+ Ke8
42. Ra8+ Kf7
43. Ra7+ Kg6
This was my plan, but now White can draw. For example: 44. d7 Rd3+ 45. Kc6 g4 46. Rxa3 Rd4 47. Ra8 h5 48. d8Q Rxd8 49. Rxd8, when, apparently almost everything draws. Who was it who first said that passed pawns should be pushed?
Now Black’s winning again after 44.. Kf5 45. d7 Rd3+ 46. Kc6 Re6. My move still wins with best play – but it’s far from trivial.
45. Kd5 g4
46. d7 Rd3+
47.. Kc6 was more challenging: Stockfish analyses 47. Kc6 f5 48. Rxa3 Rd2 49. Ra8 Kg5 50.d8Q+ Rxd8 51.Rxd8 g3 52.Kd5 Kf4 53.Ke6 h5 when Black will apparently end up with KQ v KR. It looks more natural to head towards the pawns, though. These endings with rook against three connected pawns are tremendously difficult to calculate. A slight difference in the placing of the pieces can make a bit difference to the assessment.
Now we see the reason why 47.. Kc6 was better. 48. Rxa3 no longer works because there’s a skewer coming up: 48.. Rxa3 49. d8Q Re3+ and Rd3+.
Missing the win again: 48.. Re3+ 49. Kd6 h5 should have been played.
49. Rg8+ Kh7?
I should have preferred 49.. Kh6 50. Rxg3 Rd2 51. Rxa3 Rxd7 52. Kxd7 Kg5, which is drawn.
But now White can turn the tables with 50. Rxg3 Rxg3 (50. Rd2 loses here because the black king is too far back) 51. d8Q a2 52. Qd4! (the only winning move) and this time it’s White who will have KQ v KR after eliminating the black pawns.
Now Black’s winning again.
51. Rxd8 g2
52. Rd1 h4
The winning move was 53.. a2! 54. Kf7 Kh6! (54.. h3 55. Rd3 g1N! also wins, for fans of underpromotions), the point being that after Kf6 the a-pawn will promote with check.
A half point is available again for White: 54. Rd7+ Kh6 55. Rd8 Kh5 56. Kf5 Kh4 57. Kf4. I seemed to recall seeing an endgame study which concluded like this once.
54. Kf7? h2
55. Rd4 h1=Q
56. Rg4 and White resigned
A few lessons to learn from this ending:
- Not all endings are boring
- If you have a passed pawn, push it – but if you have more than one, be careful to make the right choice
- If your opponent has a passed pawn, keep it under lock and key
- Mediocre chess is often more exciting than good chess
- Mediocre chess is often more instructive than good chess
- Practising endings is really important
- Solving endgame studies can be helpful
- It might be worth your while to learn how to play KQ v KR
- Chess is hard – and rook endings are especially hard
Here’s the complete game: