4

I was playing at my local chess club on Sunday and I had the position below as white. Me and my opponent continued for a while before finally agreeing to a draw. When I play it against bots as white sometimes I lose and sometimes I draw. For example, when I played against the lichess bot I lost, but when I played against the chess.com bot I got a draw. Is the position below a draw?


[Title "White to Move"]
[FEN "6k1/7p/6p1/5q2/6R1/5P1P/6PK/8 w - - 0 1"]
[FEN "7k/6p1/7p/5q2/6R1/5P1P/6P1/6K1 w - - 0 1"]

1. Re4 Qc5+ 2. Kh2 Qd6+ 3. Kg1 Kg8 4. Rg4 h5
5. Re4 g5 6. Rc4 Kf7 7. Re4 Kf6 8. Rc4 Kf5
9. Re4 Qd1+ 10. Kh2 Qd6+ 11. Kg1 Qb8 12. Rc4 Qb6+
13. Kh2 Ke5 14. Re4+ Kf5 15. Kh1 Kg6 16. Ra4 Kf5
17. Re4 Qb1+ 18. Kh2 Qd1 19. Rc4 Qd6+ 20. Kh1 Qd1+
21. Kh2 Qd6+ 22. Kh1 Kf6 23. Re4 Kg6 24. Rc4 Qd5
25. Re4 Qd1+ 26. Kh2 Kf6 27. Rc4 Qd6+ 28. Kg1 Qb6+
29. Kh1 Ke7 30. Kh2 h4 31. Re4+ Kf7 32. Rg4 Ke6
33. Re4+ Kd6 34. Rg4 Ke6 35. Re4+ Kd5 36. Rg4 Qb1
37. Re4 Qc1 38. Rg4 Qc7+ 39. Kg1 Ke6 40. Re4+ Kd5
41. Rg4 Qc1+ 42. Kh2 Qb1 43. Re4 Qf1 44. Rg4 Ke5
45. Re4+ Kf6 46. Rg4 Qd1 47. Re4 Kf5 48. Rg4 Ke5
49. Re4+ Kd6 50. Rg4 Ke5 51. Re4+ Kf5 52. Rg4 Qc1
53. Re4 Qa1 54. Rg4 Kf6 55. Re4 Kf7 56. Rg4 Kf6
57. Re4 Kf5 58. Rg4 Ke5 59. Re4+ Kd5 60. Rg4 Kc5
61. Re4 Kd6 62. Rg4 Kc5 63. Re4 Kd5 64. Rg4 Ke5
65. Re4+ Kd6 66. Rg4 Qc1 67. Re4 Kd7 68. Rg4 Ke7
69. Re4+ Kd7 70. Rg4 Qc5 71. Re4 Kd6
1
  • The only winning attempt that Black has should be pushing h5 g5, attack the rook on e4, play g4 to force hxg4 and then go h4. I don't know if there is a way to win from there, but anything else White can just sit and do nothing.
    – koedem
    Nov 4 at 12:03

4 Answers 4

0

Adding to Ryley's answer (which gives no strategies): Assume two pawn pairs get exchanged. Pg2 + R walking between f3 and h3 is a fortress - draw. Avoid for Black. Most else Q/RP is won, but sometimes even only in DTZ 50+! My strategy for Black would be using pawn moves only very sparingly and trying to break through with your K behind the enemy lines, i.e. Kf1/f2 with no checks possible. It doesn't matter if it costs your own pawns if you also get two. (Pf3+Ph3 is drawish again.) White should teeth-clenchingly defend the rook position somewhere on the 4th.

0

Based on this line you wrote, it's a probable draw.

When I play it against bots as white sometimes I lose and sometimes I draw

Given that bots are much stronger than the best humans, they're likely to be light years better than you, so if you manage to draw some of the time it's probable that another bot will be able to draw all of the time - and hence the position is objectively drawn.

The following observation backs up the above too: if you run Stockfish in multi-PV mode on that position, it suggests several different lines that all have the same evaluation. Then, after making its suggested move, it still suggests several different lines that all have the same evaluation. This is known to engine chess fans as a "flatline" or "horizone", and is very suggestive of fortresses. It doesn't prove a fortress per se, but it is very suggestive.

If you want to analyze the position in detail, you can also increase the half-move counter in the FEN to speed up the engine pruning.

3
  • So you think this is a draw?
    – Bob
    Nov 1 at 3:02
  • @Bob yes (15 characters).
    – Allure
    Nov 1 at 3:02
  • 1
    Can I see the lines?
    – Bob
    Nov 1 at 13:36
0

Don't listen to computers! This position can very well be drawn as the 2 pawns will probably be exchanged into a Q vs R + P ending. It is very hard to win Q vs R not to mention this! The only possibility to lose is to get checkmated. Or else, it is drawn.

2
  • May I see the computer lines?
    – Bob
    Nov 6 at 20:51
  • Computer lines are not required. Simply logic.
    – MrChessR
    Nov 6 at 23:38
-3

The Lichess analysis board evaluates it to be around -2.2 which indicates a pretty (but not totally) losing position for white.

If white plays well then it's possible to draw but unless black blunders, white will almost certainly not win.

3
  • 9
    This is not a good position to rely on engine evaluations unfortunately, because it's a possible fortress, and engines still aren't very good at detecting fortress positions,
    – Allure
    Oct 30 at 9:48
  • 2
    Stockfish 14+ NNUE evaluates this as –1.8 at depth 58 with 20336k node/sec. This means very little. Playing through various engine lines, including the principal variation, using depth = 50 for each move results in a 7-piece tablebase draw (many of these drawn positions seem to have a –1.5 evaluation if the engine doesn't see the draw yet). Oct 30 at 12:42
  • 2
    When analyzing endgames we shouldn't consider such things as "pretty lost positions". It's either lost or drawn!
    – David
    Nov 1 at 13:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.