interesting sac

[Vinvin]

3 best moves after 2h20 C2D 2.13 ghz

Analysis by HIARCS 12 MP:

1. -+ (-1.97): 1...Te1 2.Rg2 Cg5 3.f6 gxf6 4.Fxg5 Txa1 5.Ff4 Txa2 6.b3 Dg7+ 7.Rh2 Tg8 8.Fg3 Txc2 9.Fc4 Td8 10.Df4 Tb2 11.h4 Te8 12.Fd5 Tbe2
2. µ (-1.36): 1...Cxf2 2.Rxf2 Dc5+ 3.Rg2 Dxc2+ 4.De2 Txc1 5.Dxc2 Txc2+ 6.Rg3 Txb2 7.a4 Txf5 8.a5 Tg5+ 9.Rf3 Tb3+ 10.Rf2 Tgg3 11.a6 Ta3 12.Td1 Taf3+ 13.Re1 Tf8
3. µ (-1.13): 1...Dh4 2.Rg2 Cxf2 3.Fe2 Th1 4.Fe3 Txa1 5.Fxf2 Df6 6.b3 Txa2 7.Fd3 Df7

[Tony Thomas]

From what I know Xyclops is still a private engine, another engine from the same author was released few days ago. If I had to take a guess, Xyclops is stronger than Arasan considering that it is running on a single processor. I played few games against it using Romi, and Romi was brutally defeated, so far not even one draw. On the other hand, Romi is able to win/draw Arasan once in a while..

What is the other engine from the same author that was released a few days ago Tony

It's called Alderon and you can download it from Xyclops homepage..It looks as if Xyclops is a rewrite...By the way, I finally got a draw..

http://www.xyclops.com/contact.html

[Dr.Wael Deeb]

From what I know Xyclops is still a private engine, another engine from the same author was released few days ago. If I had to take a guess, Xyclops is stronger than Arasan considering that it is running on a single processor. I played few games against it using Romi, and Romi was brutally defeated, so far not even one draw. On the other hand, Romi is able to win/draw Arasan once in a while..

What is the other engine from the same author that was released a few days ago Tony

It's called Alderon and you can download it from Xyclops homepage..It looks as if Xyclops is a rewrite...By the way, I finally got a draw..

http://www.xyclops.com/contact.html

Thanks Tony,me the Great Goat bless you my friend

[jdart]

Arasan has picked up a couple wins against xyclops, this is one:

[Event "?"]
[Site "chessclub.com"]
[Date "2008.07.01"]
[Round "?"]
[White "Arasan 10.4"]
[Black "xyclOps"]
[Result "1-0"]
[ECO "C45"]
[WhiteElo "2682"]
[BlackElo "2779"]
[TimeControl "300+8"]

1. e4 e5 2. Nf3 Nc6 3. d4 exd4 4. Nxd4 Nge7 5. Nc3 Nxd4 6. Qxd4 Nc6
7. Qe3 Nb4 8. Qd2 Be7 9. Bc4 O-O 10. O-O d6 11. f4 Be6 12. Bxe6 fxe6
13. a3 Nc6 14. Qd3 Bf6 15. Be3 Kh8 16. Kh1 Qe8 17. Rf3 Rd8 18. Nb5
Qe7 19. Rh3 g6 20. Rb1 a6 21. Nc3 Rg8 22. Ne2 g5 23. Rf1 g4 24. Rh5
Rge8 25. c3 Rf8 26. a4 Rde8 27. b4 Qf7 28. Ng3 Rd8 29. Qe2 Qg8 30. b5
axb5 31. Qxb5 Rf7 32. Qxb7 Qe8 33. Qb3 Ra8 34. Bd2 Rf8 35. Qd1 Rg8
36. Rh6 Bg7 37. Rh5 Bf6 38. Qb3 Rf8 39. Qc2 e5 40. f5 Qf7 41. Rh6 Qc4
42. Qd1 Qg8 43. Rg1 Ra6 44. Be3 Raa8 45. h4 Ne7 46. Kh2 Ra6 47. Rf1
Raa8 48. Rf2 Qc4 49. Qxg4 Ng8 50. Rh5 Rf7 51. a5 Qxc3 52. Bd2 Qc4
53. Qd1 Be7 54. Rf3 Qb5 55. Rb3 Qa6 56. Kh3 Nf6 57. Rh6 Qa7 58. Qe2
Bf8 59. Rxf6 Rxf6 60. Bg5 Rf7 61. Qc4 c6 62. Nh5 h6 63. Bf6+ Kh7
64. Rg3 d5 65. Qc1 Rc8 66. Qd1 Qc7 67. Qg4 Bg7 68. Qg6+ Kg8 69. Nxg7
Rxg7 70. Bxg7 Qf7 71. Qxh6 Qxf5+ 72. exf5 Rd8 73. Bxe5+ Kf7 74. Qh7+
Ke8 75. Rg8# 1-0 {xyclOps checkmated}

[beachknight]

Hi Tony,

Has Alderon a config file or clps to run on Arena beta?

best,

[Tony Thomas]

Hi Tony,

Has Alderon a config file or clps to run on Arena beta?

best,

As far as I know it doesnt have a config, and I am not sure what clps is (commandline blah blah??).

[Eelco de Groot]

This is with a different version, it is supposed to know a little more about control of the centre, mainly indirectly by counting attacked squares so if it is working it could be called Hypermodern Centre Control I have not implemented anything new yet for pawns controlling centre squares let alone pawns occupying it. In this testposition there are no center pawns so here it should not matter hopefully and there is already the existing code and Piece Square Tables, so maybe it would not make any sense changing it or making new code especially for Pawns, most effect I expect for sliding pieces Rooks, Bishops and Queens.

Created two small new bitboards and put them in bitboard.h:

Code: Select all

const Bitboard Centre = 0x1818000000ULL;
const Bitboard ExpandedCentre = 0x3C3C3C3C0000ULL;

Now for all pieces after computing mobility the same bitboard with attacked squares is reused for counting which of those attacked squares are in the centre, e4, d4, e5, d5 or the expanded centre c3-f3, c4-f4, c5-f5, c6-f6. There are no different weights for the pieces and I just guessed these weights, so far I don't know if the computation is really worth it, it does slow the engine down again.

This is the computed Centre Control for the Queen:

Code: Select all

Bitboard b = p.queen_attacks(s);
    ei.attackedBy[us][QUEEN] |= b;

    // King attack
    Bitboard bb = (b & ei.attackZone[us]);
    if(bb) {
      ei.attackCount[us] += count_1s_max_15(bb);
      ei.attackWeight[us] += QueenAttackWeight;
      Bitboard bbb = (b & ei.attackedBy[them][KING]);
      if(bbb) ei.attacked[us] += count_1s_max_15(bbb);
    }
    
    // Mobility
    int mob = count_1s(b & ~p.pieces_of_color(us) & ~ei.attacked_by(them));
    ei.mgMobility += Sign[us] * MidgameQueenMobilityBonus[mob];
    ei.egMobility += Sign[us] * EndgameQueenMobilityBonus[mob];

    // Centre Control
    int control = count_1s_max_15(b & Centre);
    ei.mgCentreControl += Value(Sign[us] * control * 8);
    ei.egCentreControl += Value(Sign[us] * control * 6);
    control = count_1s_max_15(b & ExpandedCentre);
    ei.mgCentreControl += Value(Sign[us] * control * 4);
    ei.egCentreControl += Value(Sign[us] * control * 3);
  }

Tord will maybe also see a difference in above code how ei.attackCount is computed in Ancalagon:

Code: Select all

  // attackCount[color] is the number of pieces of the given color which
  // attack a square adjacent to the enemy king.
  // [EdG: This definition has changed in Ancalagon; now attackCount[color]
  //  counts the number of attacks on attackZone[color], not the attackers.]
  int attackCount[2];

Maybe the bonuses for centre control have to be much lower but this is just the very first test of the code:

[d]5r1k/4q1pp/8/5P2/4n3/5Q1P/PPP2P2/R1Br1BK1 b - -

Engine: Ancalagon 2.0 Beta 006j settings Build 70 (Athlon 2009 MHz, 256 MB)
by Tord Romstad, Eelco de Groot

2.00 0:00 -1.80 1...Rfd8 2.Qh5 (255) 1

2.00 0:00 -1.74 1...Rd5 2.Bd3 (426) 1

3.00 0:00 -1.98 1...Rd5 2.Bd3 Rd4 (878) 3

3.00 0:00 -1.50 1...Rfd8 2.Qf4 Kg8 (1.438) 6

3.00 0:00 -0.70 1...Re1 2.Qa3 Qxa3 3.bxa3 Rxf5 (1.879) 7

4.00 0:00 -0.70 1...Re1 2.Qa3 Qxa3 3.bxa3 Rxf5 (3.047) 12

5.00 0:00 -0.82 1...Re1 2.Qa3 Qxa3 3.bxa3 Rxf5 4.f4 (5.752) 23

5.00 0:00 -0.43 1...Rfd8 2.Qf4 Nd2 3.Bxd2 Rxa1 4.a4 (9.435) 35

6.00 0:00 -0.13 1...Rfd8 2.b4 Qh4 3.Bb2 R1d2 4.Qg2 (16.149) 57

7.00 0:00 +0.15 1...Rfd8 2.b4 Qh4 3.Bb2 R1d2 4.Be2 Rxc2 (67.459) 154

8.00 0:00 +0.09 1...Rfd8 2.b4 Qh4 3.Bb2 R1d2 4.Be2 Rxc2
5.Be5 (109.576) 194

9.00 0:01 +0.27 1...Rfd8 2.a4 Qh4 3.Qg2 R8d5 4.f6 Rd7
5.a5 gxf6 (386.496) 291

10.01 0:02 +0.17 1...Rfd8 2.b4 Qxb4 3.Kg2 Qa4 4.Bd3 Qd4
5.Bb2 Rxd3 6.Bxd4 Rxf3 7.Kxf3 Rxd4 (847.394) 324

11.01 0:06 -0.35 1...Rfd8 2.b3 Qh4 3.Bb2 R1d2 4.Qg2 Rg8
5.Re1 Ng5 6.Kh1 Rxf2 (2.346.274) 353

11.02 0:09 +0.54 1...Re1 2.b4 Nd2 3.Bxd2 Rxa1 4.b5 Rxa2
5.Qb3 Rfa8 6.Qe6 Qb7 7.Bg2 (3.373.172) 360

12.01 0:26 +0.72 1...Re1 2.Kg2 Nd2 3.Bxd2 Rxa1 4.Bc3 Rc1
5.Qf4 Qb7+ 6.Kg1 Ra1 7.Qg4 Rxa2
8.Qh5 (9.593.866) 365

13.01 0:45 +0.56 1...Re1 2.Kg2 Nd2 3.Bxd2 Rxa1 4.c4 Rxa2
5.Bc3 Ra1 6.b4 Ra2 7.Qf4 Qb7+ 8.Kh2 Kg8 (16.667.008) 364

13.02 1:34 +2.33 1...Qh4 2.b4 Nxf2 3.Qxf2 Qxf2+
4.Kxf2 Rxf5+ 5.Ke2 Rfxf1 6.b5 Rxc1
7.Rxc1 Rxc1 8.Kd3 Rd1+ 9.Ke4 Rd2
10.b6 Rxc2 (34.097.288) 361

14.01 1:51 +2.54 1...Qh4 2.b4 Nxf2 3.Qxf2 Qxf2+
4.Kxf2 Rxf5+ 5.Ke2 Rfxf1 6.b5 Rxc1
7.Rxc1 Rxc1 8.Kd3 Rh1 9.Kd4 Rxh3
10.b6 Kg8 (40.352.965) 361

15.01 2:45 +2.74 1...Qh4 2.Kg2 Nxf2 3.Be2 Rh1 4.Bf4 Rxa1
5.Qxf2 Qe7 6.Bd3 Qb7+ 7.Kg3 Rxa2
8.Kh2 Ra1 9.Qg2 Qxb2 (60.133.968) 362

16.01 4:08 +2.58 1...Qh4 2.Kg2 Nxf2 3.Be2 Rh1 4.Bf4 Rxa1
5.Qxf2 Qe7 6.Bd3 Qb7+ 7.Kg3 Rxa2
8.Qf3 Qxb2 9.Qd5 Re8 10.Qf7 (90.227.800) 362

17.01 11:27 +2.74 1...Qh4 2.Kg2 Nxf2 3.Be3 Rxa1 4.Bxf2 Qd8
5.h4 Rxa2 6.Bd3 Rxb2 7.Bc5 Re8 8.Qc6 Rb1
9.Kh3 Kg8 10.Bc4+ Kh8 (248.493.777) 361

18.01 22:56 +2.17 1...Qh4 2.Kg2 Nxf2 3.Be3 Rxa1 4.Bxf2 Qg5+
5.Qg4 Qh6 6.Bd3 Rxa2 7.Qh4 Qxh4
8.Bxh4 Rxb2 9.Kf3 Rc8 10.Bg3 Kg8
11.Be5 Rb1 (489.596.041) 355

19.01 64:08 +2.00 1...Qh4 2.Kg2 Nxf2 3.Be3 Rxa1 4.Bxf2 Qg5+
5.Qg4 Qh6 6.Bd3 Rxa2 7.Qh4 Qxh4
8.Bxh4 Rxb2 9.Kf3 Rc8 10.Bg3 Kg8
11.Be5 Rb1 12.Ke4 (1.367.412.836) 355

At depth 19 there is no switch to 1... Re1, at the moment Ancalagon is looking at 1... Nxf2 at depth 19 after circa 100 minutes into the search.

[beachknight]

Thanks Tony.

Alderon does not run on Arena beta 1.99 5 on my quad.

No problem in console mode, though.