Absolute ELO scale

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

kbhearn
Posts: 411
Joined: Thu Dec 30, 2010 4:48 am

Re: Absolute ELO scale

Post by kbhearn »

in principle it should still be possible without forced captures

1) win all the nonpawn material you can and blockade all the enemy pawns you can
2) force the enemy king to capture any of your pawns you need out of the way
3) arrange your pieces to smother your own king
4) force the enemy king into a stalemated position where it will support its pawn delivering checkmate and unblock the pawn that is to reach there - waste moves til mate is delivered
User avatar
Guenther
Posts: 4605
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: Absolute ELO scale

Post by Guenther »

cdani wrote:Here they are:

Randscacs:
www.andscacs.com/randscacs089025.zip
Able to play arbitrarily long games, and also can play depth 1 random moves. Is updated with the current development version of Andscacs.

Andworst:
www.andscacs.com/andworst-0.3.zip
Able to play arbitrarily long games.
Daniel, if it's not too much hassle would you mind creating no popcnt versions of both?
User avatar
hgm
Posts: 27787
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: Absolute ELO scale

Post by hgm »

kbhearn wrote:in principle it should still be possible without forced captures

1) win all the nonpawn material you can and blockade all the enemy pawns you can
2) force the enemy king to capture any of your pawns you need out of the way
3) arrange your pieces to smother your own king
4) force the enemy king into a stalemated position where it will support its pawn delivering checkmate and unblock the pawn that is to reach there - waste moves til mate is delivered
This might be quite hard. It will hardly be possible to force the opponent to do anything if he still has non-Pawns. To checkmate with a Pawn, that Pawn would have to be protected, and if it is protected by another Pawn, that Pawn would have to be blocked. But that means the Pawn that should be forced to mate you will be able to capture the blocker, so it is not really forced.

You cannot force a Pawn to mate you on a promotion move, because he can always choose a piece that doesn't check.

This might do it, though:
[d]8/2Q5/k7/2R5/8/2pp4/p1p4P/2K2B2 w - - 0 1
kbhearn
Posts: 411
Joined: Thu Dec 30, 2010 4:48 am

Re: Absolute ELO scale

Post by kbhearn »

I agree it'd require a lot more extra material than it does to deliver a mate yourself, but most games that won't be a problem, you'll just run extra promotions. A relatively simple example of a position that if you aimed for it would finish well:

[pgn]
[FEN "8/8/1p6/1B6/3Q3R/k7/8/K7 w - - 0 1"]

1.Bd3 b5 2.Bb1 b4 3.Qd1 b3 4.Rg4 b2# 0-1
[/pgn]

[pgn]
[FEN "8/8/1p6/1B6/3Q3R/k7/8/K7 w - - 0 1"]

1.Bd3 Kb3 2.Bb1 Ka3 3.Qd1 b5 4.Rg4 b4 5.Rh4 b3 6.Rg4 b2# 0-1
[/pgn]

I'm not sure what minimal material to actually force mate would be - would depend on what pawns the enemy has left at least, but certainly against a player that's trying to lose by giving away material such a position could be consistently forced.
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Absolute ELO scale

Post by cdani »

Guenther wrote:
cdani wrote:Here they are:

Randscacs:
www.andscacs.com/randscacs089025.zip
Able to play arbitrarily long games, and also can play depth 1 random moves. Is updated with the current development version of Andscacs.

Andworst:
www.andscacs.com/andworst-0.3.zip
Able to play arbitrarily long games.
Daniel, if it's not too much hassle would you mind creating no popcnt versions of both?
Done. Download the same zip files and they are there.
User avatar
Guenther
Posts: 4605
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: Absolute ELO scale

Post by Guenther »

cdani wrote:
Guenther wrote:
cdani wrote:Here they are:

Randscacs:
www.andscacs.com/randscacs089025.zip
Able to play arbitrarily long games, and also can play depth 1 random moves. Is updated with the current development version of Andscacs.

Andworst:
www.andscacs.com/andworst-0.3.zip
Able to play arbitrarily long games.
Daniel, if it's not too much hassle would you mind creating no popcnt versions of both?
Done. Download the same zip files and they are there.
Thanks Daniel! Very helpful as always. Happy X-mas in advance :)

I will try to lose a game against Andworst, but after Kevins post I am even more sceptical I can reach that goal in a normal time control.
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Absolute ELO scale

Post by cdani »

Guenther wrote: Thanks Daniel! Very helpful as always. Happy X-mas in advance :)

I will try to lose a game against Andworst, but after Kevins post I am even more sceptical I can reach that goal in a normal time control.
Happy Christmas also!!
Good luck trying to lose! :-)
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: Absolute ELO scale

Post by Sven »

cdani wrote:
Guenther wrote: Thanks Daniel! Very helpful as always. Happy X-mas in advance :)

I will try to lose a game against Andworst, but after Kevins post I am even more sceptical I can reach that goal in a normal time control.
Happy Christmas also!!
Good luck trying to lose! :-)
I have played a 40/10 game against Andworst. I haven't lost but "managed" to draw:
[pgn][Event "Computer Chess Game"]
[Site "PRIMUS"]
[Date "2016.12.21"]
[Round "-"]
[White "sven"]
[Black "Andworst -0.3"]
[Result "1/2-1/2"]
[TimeControl "40/600"]
[Annotator "1... -0.71"]

1. f3 g5 {-0.71/21 11} 2. Kf2 Na6 {-0.31/21 9} 3. Ke3 Nb8 {+0.01/20 8} 4.
Kd4 Na6 {+0.01/27 9} 5. Kd5 Nc5 {-0.65/27 30} 6. c3 Ne4 {-2.18/23 29} 7.
Qb3 f6 {-2.64/22 29} 8. Kd4 Ng3 {-2.76/21 14} 9. Qf7+ Kxf7 {+8.04/4} 10. e3
Ke6 {+7.36/21 8} 11. Bc4+ Kf5 {+7.27/22 11} 12. Be6+ Kxe6 {+11.60/22 2.9}
13. Ne2 Kd6 {+10.39/22 8} 14. b3 Bh6 {+9.56/21 9} 15. Ba3+ Ke6
{+10.23/22 9} 16. Bd6 Kf5 {+10.30/23 19} 17. Nf4 Nf1 {-99.98/1} 18. Nh5 Ng3
{+0.01/30 8} 19. Na3 Nf1 {+0.01/31 11} 20. d3 Nxh2 {+13.18/31 12} 21. Nc4
Qe8 {+12.30/18 16} 22. e4+ Kg6 {+12.51/21 9} 23. Ne5+ Kxh5 {-99.98/1} 24.
Nc6 Qg6 {-99.98/1} 25. g4+ Kh4 {-99.98/1} 26. Bg3+ Kh3 {-99.92/127 14} 27.
Rxh2+ Kxg3 {+17.30/4} 28. Rg1+ Kxf3 {+0.01/46 32} 29. Rf2+ Kxf2 {+26.70/4}
30. Nb8 Qf5 {+26.74/21 27} 31. c4 Ke2 {+18.72/22 31} 32. Re1+ Kf2
{+0.01/48 17} 33. Rf1+ Ke2 {+0.00/52 14} 34. Re1+ Kf2 {+0.01/50 15} 35.
Rf1+ Ke2 {+0.01/50 14} 36. Rf2+ Kd1 {+17.67/22 29} 37. Rd2+ Kc1
{+28.30/24 17} 38. Kc3 Kb1 {+0.01/43 32} 39. Rd1+ Kxa2 {+12.56/4} 40. Ra1+
Kxa1 {+21.92/4} 41. Kd4 a6 {+26.67/21 34} 42. e5 e6 {+29.68/22 33} 43. Kc5
Bg7 {+30.18/23 32} 44. d4 a5 {+33.40/20 19} 45. Nc6 b5 {+42.32/20 31} 46.
Nb4 Qd3 {+30.02/20 30} 47. Nd5 Qf5 {+0.01/41 9} 48. Nb4 Qd3 {+0.01/40 11}
49. Nc2+ Kb1 {+99.91/42 9} 50. b4 Qa3 {+91.30/21 28} 51. Na1 Qd3
{+0.01/44 9} 52. Nc2 Qa3 {+0.01/44 8} 53. Na1 Qd3 {+0.01/45 9} 54. Nc2
{XBoard adjudication: repetition draw} 1/2-1/2[/pgn]
User avatar
Guenther
Posts: 4605
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: Absolute ELO scale

Post by Guenther »

cdani wrote:
Guenther wrote: Thanks Daniel! Very helpful as always. Happy X-mas in advance :)

I will try to lose a game against Andworst, but after Kevins post I am even more sceptical I can reach that goal in a normal time control.
Happy Christmas also!!
Good luck trying to lose! :-)
My first attempt for warming up. I admit that I had no real plan yet.
At least I could sign a morally loss by having just K vs. K,B,N in the end
hoping for a bug of Andworst, but it simply drew that.

Obviously I let need it underpromoting at least 4 pawns and also having my K already in the corner when doing so.

BTW there was one surprising move 19...Nxa6? which ruined my strategy
and seems a bit strange when playing for the worst move?
OTH I realized then that I cannot predict its moves ;-)

[pgn]
[Event "RWBC"]
[Site "CAPPUCCINO"]
[Date "2016.12.21"]
[Round "-"]
[White "Guenther Simon"]
[Black "Andworst_03-64"]
[Result "*"]
[TimeControl "40/300"]
[Annotator "1... -0.57"]

1. f3 g5 {-0.57/18 7} 2. h4 Nf6 {-1.61/23 14} 3. hxg5 Bh6 {-5.52/23 14} 4.
Kf2 O-O {-7.13/21 14} 5. Rxh6 e6 {-9.53/22 13} 6. gxf6 Qe8 {-99.90/24 13}
7. d3 e5 {-99.92/30 13} 8. e4 c5 {-99.92/29 12} 9. Ne2 a6 {-99.92/29 12}
10. Ng3 b5 {-99.92/30 12} 11. Nf5 Bb7 {-99.92/27 11} 12. Be2 Qe7
{-99.94/124 11} 13. Nxe7+ Kh8 {-99.96/4} 14. Nc3 Bd5 {-99.96/127 1.0} 15.
Ncxd5 Rc8 {-99.96/127 1.1} 16. a4 bxa4 {-99.96/127 0.5} 17. Rxa4 Rc7
{-99.96/127 0.3} 18. Be3 c4 {-99.96/127 0.6} 19. Rxa6 Nxa6 {-99.96/127 1.5}
20. c3 cxd3 {-99.96/127 2.9} 21. Qa4 Nb4 {-99.96/127 0.7} 22. Nxb4 d2
{-99.96/127 0.6} 23. Bc4 d6 {-99.96/127 0.7} 24. Be6 Ra5 {-99.98/1} 25.
Rxh7+ Kxh7 {-99.96/4} 26. Nbd5 Rxc3 {-99.96/127 0.5} 27. bxc3 Ra8
{-99.96/127 0.4} 28. Bh6 Kh8 {-99.96/127 0.4} 29. Bg7+ Kh7 {-99.98/1} 30.
Nb6 d5 {-99.98/1} 31. Nxa8 d4 {-99.98/1} 32. c4 d3 {-99.98/1} 33. Qc2 dxc2
{-99.98/1} 34. Ng6 d1=B {-99.96/127 0.3} 35. Nc7 Kg8 {-99.96/127 0.3} 36.
Bxf7+ Kh7 {-99.98/1} 37. f4 Be2 {-99.98/1} 38. Ne6 Bd1 {-99.98/1} 39. Nd4
Be2 {-99.98/1} 40. Nf3 Bd1 {-99.98/1} 41. g4 Be2 {-99.98/1} 42. Ngh4 Bd1
{-99.98/1} 43. Ng2 Be2 {-99.98/1} 44. Ne3 Bd1 {-99.98/1} 45. Nd4 Bf3
{-99.96/127 0.3} 46. Nb3 c1=B {-99.94/127 2.6} 47. Nd2 Bb2 {-99.96/127 0.4}
48. c5 Bxe4 {-99.96/127 0.3} 49. c6 Ba3 {-99.96/127 0.3} 50. c7 Bg2
{-99.96/127 0.4} 51. c8=N Be4 {-99.96/127 0.5} 52. Bd5 Bg6 {-99.94/127 7}
53. f5 Be7 {-99.94/127 4} 54. Nd6 Bxd6 {-99.94/127 5} 55. Ne4 Ba3
{-99.98/1 0.1} 56. Bg8+ Kxg8 {-9.10/4} 57. Ng5 Bxf5 {-99.94/127 2.0} 58.
gxf5 Bb2 {-99.94/127 1.0} 59. Nd5 e4 {-99.98/1} 60. Ke3 Ba1 {-99.98/1} 61.
Ne6 Bd4+ {-99.90/127 59} 62. Ke2 Kh7 {-99.96/127 0.3} 63. f7 Bg1 {-99.98/1}
64. Ne3 Bh2 {-99.96/127 0.3} 65. Nf4 Bg1 {-99.96/127 0.3} 66. Bh6 Kh8
{-99.96/127 0.2} 67. Bg5 Bxe3 {-99.94/127 1.0} 68. Nd5 Bd2 {-99.96/127 0.2}
69. Bf6+ Kh7 {-99.96/4 0.1} 70. Nf4 Kh6 {-99.96/127 0.2} 71. Bg5+ Kh7
{-99.92/127 5} 72. Bf6 Ba5 {-99.96/127 0.3} 73. Be5 e3 {-99.96/127 0.2} 74.
Nh3 Be1 {-99.96/127 0.2} 75. Ng5+ Kh6 {-11.15/4} 76. Bc3 Bd2
{-99.90/127 1:31} 77. f8=B+ Kh5 {-68.44/39 1:15} 78. Nf3 Be1
{-99.80/29 1:01} 79. Bh6 Bh4 {-99.84/28 48} 80. Bg5 Be1 {-99.84/28 37} 81.
Bh4 Kh6 {-99.82/25 16} 82. f6 Bf2 {-99.88/26 16} 83. Be1 Kh5 {-99.88/26 15}
84. Ne5 Kh6 {-99.90/31 15} 85. f7 Bxe1 {-99.94/127 0.8} 86. Bf6 Bd2
{-99.96/127 0.3} 87. f8=Q+ Kh7 {-99.98/1} 88. Qf7+ Kh6 {-99.98/1} 89. Bg7+
Kg5 {-99.96/127 0.2} 90. Nf3+ Kg4 {-74.84/4} 91. Qg6+ Kf4 {-99.98/1} 92.
Ne1 Bc1 {-99.98/1} 93. Kd3 e2 {-99.96/127 0.3} 94. Bd4 Ba3 {-99.96/127 0.2}
95. Qe4+ Kg3 {-99.96/127 0.2} 96. Ke3 Kh2 {-99.98/1} 97. Kf2 Bc1 {-99.98/1}
98. Be3 Bb2 {-99.98/1} 99. Qg4 Ba1 {-99.98/1} 100. Bd4 Bb2 {-99.98/1} 101.
Nf3+ Kh1 {-99.98/1} 102. Qc8 Ba1 {-99.98/1} 103. Qc3 Bb2 {-99.98/1} 104.
Ne1 Ba1 {-99.98/1} 105. Qb2 Kh2 {-99.96/127 0.2} 106. Nf3+ Kh1
{-99.96/127 0.2} 107. Nh2 Kxh2 {-99.96/127 0.2} 108. Be5+ Kh1
{-99.98/127 0.2} 109. Bh2 e1=N {-99.96/127 0.1} 110. Kf1 Nd3 {-99.98/1}
111. Bg3 Nc1 {-99.98/1} 112. Ke1 Na2 {-99.98/1} 113. Kd2 Nc3
{-99.88/127 59} 114. Qb7+ Ne4+ {-99.92/89 48} 115. Kc2 Bc3 {-99.90/31 39}
116. Be5 Be1 {-99.92/57 32} 117. Qd5 Bd2 {-99.92/36 14} 118. Bd4 Kg2
{-99.90/27 25} 119. Kb1 Kf1 {-99.90/26 20} 120. Be3 Ng3 {-99.94/127 5} 121.
Qf3+ Ke1 {-99.96/4} 122. Qf2+ Kd1 {-99.98/1} 123. Ka1 Nh1 {-99.98/1} 124.
Kb2 Ng3 {-99.98/1 0.1} 125. Qf1+ Be1 {-99.96/127 0.2} 126. Bd2 Ne2
{-99.98/1} 127. Qf3 Bg3 {-99.94/127 1.0} 128. Bf4 Bh2 {-99.98/1} 129. Qd3+
Ke1 {-70.77/4} 130. Bg3+ Kf1 {-99.94/127 0.8} 131. Qf3+ Kg1 {-99.96/4} 132.
Qf4 Nc1 {-99.96/127 0.3} 133. Qf3 Nb3 {-99.96/127 0.2} 134. Qf2+ Kh1
{-99.98/1} 135. Be5 Na1 {-99.98/1} 136. Qg1+ Kxg1 {+0.00/127 9} 137. Kb1
Nb3 {+0.00/44 23} 138. Ka2 Bxe5 {+0.00/42 21} 139. Kb1 Nc1 {+0.00/51 19}
140. Kc2 Bd4 {+0.00/51 17} 141. Kb1 Bf2 {+0.00/52 17} 142. Ka1 Na2
{+0.00/52 15}
*
[/pgn]