v2.9, August 2021
Tree searching improvements
Positional evaluation improvements
The Nalimov endgame tablebases are now used when using the UCI protocol (if you can't use the Winboard protocol)
50 elo points of improvement (more or less)
http://technochess.free.fr
KnightX 2.9 Winboard engine released
Moderators: hgm, Rebel, chrisw
-
chrjly
- Posts: 37
- Joined: Wed Feb 03, 2021 10:45 pm
- Full name: Christophe Jolly
-
Graham Banks
- Posts: 41491
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
-
AdminX
- Posts: 6340
- Joined: Mon Mar 13, 2006 2:34 pm
- Location: Acworth, GA
Re: KnightX 2.9 Winboard engine released
Webpage shows KnightX 2.8. Anyway I downloaded the latest version (2.8). Here was a nice takedown of Rebel 13, KnightX playing with the black pieces.
[pgn]
[Event "Banksia game"]
[Site "?"]
[Date "2021.09.11"]
[Round "?"]
[White "Rebel 13"]
[Black "KnightX 2.8"]
[Result "0-1"]
[ECO "A10"]
[PlyCount "136"]
[EventDate "2021.??.??"]
[TimeControl "40/7200+1:0"]
1. c4 {1.00/1 26} g6 {0.00/2 10 1; A10: English opening} 2. g3 {1.00/1 10} Bg7
{0.18/24 240917 240422313} 3. Bg2 {1.00/1 17} e5 {0.06/26 232917 181666247} 4.
Nc3 {1.00/1 17} d6 {0.00/28 225211 193727278} 5. e4 {1.00/1 17} Ne7 {0.03/28
217714 254286406} 6. Nge2 {1.00/1 54} c5 {0.00/28 210506 213850996} 7. O-O {
0.11/18 147177 298229516} O-O {0.00/29 203508 128489009} 8. f4 {0.10/18 266935
625362667} Nbc6 {-0.18/26 196818 191652489} 9. d3 {0.13/18 35600 505795221} Be6
{-0.12/24 190212 31589879. A36: Symmetrical English vs ...g6: 4 Bg2 Bg7.} 10.
a3 {0.16/16 275654 366926117} Qd7 {-0.24/22 [#] 183917 106459544} 11. fxe5 $146
{-0.78/21 134035 369963496} (11. Rb1 $11 {-0.22/24 remains equal.}) ({
Predecessor:} 11. Rb1 Bg4 12. Qc2 a5 13. Nd5 Rae8 14. fxe5 Nxd5 15. cxd5 Nxe5
16. Nf4 Qc8 17. b4 cxb4 18. Qxc8 Rxc8 19. axb4 g5 20. h3 gxf4 21. hxg4 fxg3 22.
Bb2 a4 23. Ra1 b5 {0-1 (23) Gomez Gonzalez,S (1432)-Aguero Centurio,A (1842)
Asuncion 2017}) 11... dxe5 {-0.15/29 211998 130567741} (11... Nxe5 $17 {
-0.78/21} 12. Bf4 a6) 12. Nd5 $11 {0.16/18 288036 429540579} Bg4 {-0.15/23
170806 94659511} 13. Qc2 {0.19/18 267196 405470298} f6 {0.23/28 165109
107925703} (13... Bxe2 $15 {-0.31/30} 14. Qxe2 Nd4) 14. Be3 {0.25/18 390858
429963860} b6 {-0.24/23 159715 39625048} 15. Rae1 {0.00/29 51019 288709337} (
15. Nec3 $1 $14 {0.49/26}) 15... Rab8 {-0.12/21 154407 86504813} 16. Qd2 {
0.36/18 173652 292245612} Nd4 {-0.18/24 149307 106955046} 17. Nxd4 {0.33/21
214053 525988759} cxd4 {0.00/26 144310 151195549} 18. Bh6 {0.27/19 14562
342813287} Rf7 {0.00/26 139508 110700958} 19. Bxg7 {0.37/20 389498 1011119591}
Kxg7 {0.00/27 134914 109689422} 20. Rf2 {0.35/19 108256 363190331} Rc8 {
-0.06/24 130518 66006315} 21. Ref1 {0.33/18 209086 396693140} Rcf8 {-0.06/24
145793 155116388} 22. b3 {0.36/19 224315 565975577} Nxd5 {-0.03/25 121315
125160006} 23. exd5 {0.47/18 52562 127646955} Bh3 {-0.06/28 117319 106133696}
24. Bxh3 {0.62/17 172408 443455457} Qxh3 {-0.06/31 113410 123302970} 25. a4 {
0.53/18 144599 592714204} Qd7 {0.00/29 109707 126393915} 26. Re2 {0.52/19
319713 735593873} Qd6 {-0.18/32 106105 110201123} 27. b4 {0.45/19 319055
862263539} a6 {-0.18/27 102613 94765803} 28. Rb1 {0.45/17 157502 235158673} b5
{-0.33/26 99209 99099049} 29. axb5 {0.23/18 229242 438254696} axb5 {-0.33/27
95917 98607711} 30. Qa2 $1 {0.60/17 58609 384849189. next c5 is good for White.
} bxc4 {-0.33/27 92708 88236715} 31. dxc4 {0.21/17 294948 1029088535. Strongly
threatening c5.} Rd8 {-0.33/26 89707 74919363} 32. Qb3 {0.07/18 171045
380097552} Rb7 {-0.33/22 86716 94878704} 33. Kg2 {-0.45/31 154943 439920921} (
33. Ree1 $11 {0.00/31}) 33... Rc8 {-0.39/20 83911 67889999} 34. Rc2 {0.00/18
208914 557713546} Rd7 {-0.30/21 81114 68056877} 35. Qf3 {-0.05/18 370927
794668441} f5 {-0.51/23 78418 53113076} 36. Qd3 {-0.24/18 205491 444252154} Kg8
{-0.45/23 75818 71456745} 37. h3 {-0.05/17 186506 442166473} Ra8 {-0.75/21
73315 54261618} 38. Kh2 {-1.27/23 144722 312021263} ({Better is} 38. Rcb2 $15 {
-0.55/27}) 38... Rda7 {-0.10/26 70915 57457912} (38... Rf7 $17 {-1.27/23} 39.
Rcb2 Rfa7) 39. Rb3 $2 {-4.91/24 473725 1302527395. White does not recover from
this.} (39. b5 $1 $11 {-0.10/26}) 39... Ra1 $19 {-1.23/21 68616 41269414} 40.
Qd2 {-0.92/18 141487 355528099} Qf6 {-1.92/21 66409 42472844} 41. b5 {-1.10/17
67 158468974} e4 {-2.16/25 456308 166328255} 42. b6 {-2.14/21 134108 1150774810
} e3 {-5.01/24 289010 193270583} 43. Qg2 {-2.16/20 61 724767716} f4 {-5.28/23
279409 93005855. ...f3 is the strong threat.} 44. b7 {-4.44/20 231266
1039097504} Rb8 {-5.91/24 270111 240217722. aiming for ...f3.} 45. Qf3 {
-4.78/20 107300 845714013} Qf5 {-6.18/23 261115 169514646} 46. Rcb2 {-4.54/20
40 678103234} d3 {-7.89/22 252524 200098911} 47. Rb1 {-5.72/20 32590 639289879}
Rxb1 {-12.51/22 317829 341418001} 48. Rxb1 {-6.32/16 23342 61972793} e2 {
-12.60/22 233525 164409421} 49. gxf4 {-9.69/20 232636 604636548} Rxb7 $3 {
-14.16/23 [#] 225807 166147447. Deflection, Promotion} 50. Rxb7 {-5.65/22
Endgame 676430 630141146} (50. Rh1 Re7) 50... e1=Q {-14.61/16 327412 35887805.
Black mates.} 51. Rb8+ {-11.89/18 134082 432762298} Kg7 {-22.29/16 311118
42028160} 52. Rb2 {-12.64/21 24 919224087} d2 {-13.83/13 295619 13395012} 53.
Qc3+ {-16.39/23 41 1058504920} Qf6 {M+39/13 67985 52253859} 54. Qxd2 {-12.96/
20 244846 1062835892} Qxd2+ {M+38/10 1380 293919} 55. Rxd2 {-13.06/17 15482
77544623} Qxf4+ {[#] M+37/10 253 137999. Double Attack} 56. Kg2 {-20.40/19
303014 1352169080} Qxd2+ {M+38/10 2365 1154877} 57. Kg3 {-22.72/19 346947
1379513881} g5 {M+16/10 4323 4257866} 58. d6 {-327.45/19 414020 1716934798} g4
{M+11/10 6376 9143362} 59. hxg4 {-14.46/19 192854 858704255} Qxd6+ {M+10/10
1571 626710} 60. Kf3 {-17.77/19 356887 1732545336} Qd3+ {M+9/10 638 480396} 61.
Kf4 {-327.37/19 249245 1205677681} Qxc4+ {M+8/10 279 94160} 62. Kf3 {-327.35/
18 94689 481214270} Kg6 {10 M+7/10 302} 63. Ke3 {-327.39/15 9410 46580286} Kg5
{10 M+6/10 285} 64. Kd2 {-327.51/13 1240 5767535} Kxg4 {10 M+5/10 302} 65. Ke3
{-327.53/12 146 622320} Kg3 {10 M+4/10 285} 66. Kd2 {-17.28/7 25 21651} Kf3 {
10 M+3/10 325} 67. Kd1 {-327.65/3 37 1193} Ke3 {10 M+2/10 310} 68. Ke1 {
-327.67/2 17 328} Qc1# {10 mate M+1/10 292. Weighted Error Value: White=0.66/
Black=0.12} 0-1
[/pgn]
PS: If this is a typo please let me know so I can upgrade the version number. The release notes do show 2.9 on the webpage.
Thanks
UPDATE: I modded the 2.8 link and changed it to 2.9 and the Download came down. So I will now upgrade. http://technochess.free.fr/Archive/Knightx29_64.zip
[pgn]
[Event "Banksia game"]
[Site "?"]
[Date "2021.09.11"]
[Round "?"]
[White "Rebel 13"]
[Black "KnightX 2.8"]
[Result "0-1"]
[ECO "A10"]
[PlyCount "136"]
[EventDate "2021.??.??"]
[TimeControl "40/7200+1:0"]
1. c4 {1.00/1 26} g6 {0.00/2 10 1; A10: English opening} 2. g3 {1.00/1 10} Bg7
{0.18/24 240917 240422313} 3. Bg2 {1.00/1 17} e5 {0.06/26 232917 181666247} 4.
Nc3 {1.00/1 17} d6 {0.00/28 225211 193727278} 5. e4 {1.00/1 17} Ne7 {0.03/28
217714 254286406} 6. Nge2 {1.00/1 54} c5 {0.00/28 210506 213850996} 7. O-O {
0.11/18 147177 298229516} O-O {0.00/29 203508 128489009} 8. f4 {0.10/18 266935
625362667} Nbc6 {-0.18/26 196818 191652489} 9. d3 {0.13/18 35600 505795221} Be6
{-0.12/24 190212 31589879. A36: Symmetrical English vs ...g6: 4 Bg2 Bg7.} 10.
a3 {0.16/16 275654 366926117} Qd7 {-0.24/22 [#] 183917 106459544} 11. fxe5 $146
{-0.78/21 134035 369963496} (11. Rb1 $11 {-0.22/24 remains equal.}) ({
Predecessor:} 11. Rb1 Bg4 12. Qc2 a5 13. Nd5 Rae8 14. fxe5 Nxd5 15. cxd5 Nxe5
16. Nf4 Qc8 17. b4 cxb4 18. Qxc8 Rxc8 19. axb4 g5 20. h3 gxf4 21. hxg4 fxg3 22.
Bb2 a4 23. Ra1 b5 {0-1 (23) Gomez Gonzalez,S (1432)-Aguero Centurio,A (1842)
Asuncion 2017}) 11... dxe5 {-0.15/29 211998 130567741} (11... Nxe5 $17 {
-0.78/21} 12. Bf4 a6) 12. Nd5 $11 {0.16/18 288036 429540579} Bg4 {-0.15/23
170806 94659511} 13. Qc2 {0.19/18 267196 405470298} f6 {0.23/28 165109
107925703} (13... Bxe2 $15 {-0.31/30} 14. Qxe2 Nd4) 14. Be3 {0.25/18 390858
429963860} b6 {-0.24/23 159715 39625048} 15. Rae1 {0.00/29 51019 288709337} (
15. Nec3 $1 $14 {0.49/26}) 15... Rab8 {-0.12/21 154407 86504813} 16. Qd2 {
0.36/18 173652 292245612} Nd4 {-0.18/24 149307 106955046} 17. Nxd4 {0.33/21
214053 525988759} cxd4 {0.00/26 144310 151195549} 18. Bh6 {0.27/19 14562
342813287} Rf7 {0.00/26 139508 110700958} 19. Bxg7 {0.37/20 389498 1011119591}
Kxg7 {0.00/27 134914 109689422} 20. Rf2 {0.35/19 108256 363190331} Rc8 {
-0.06/24 130518 66006315} 21. Ref1 {0.33/18 209086 396693140} Rcf8 {-0.06/24
145793 155116388} 22. b3 {0.36/19 224315 565975577} Nxd5 {-0.03/25 121315
125160006} 23. exd5 {0.47/18 52562 127646955} Bh3 {-0.06/28 117319 106133696}
24. Bxh3 {0.62/17 172408 443455457} Qxh3 {-0.06/31 113410 123302970} 25. a4 {
0.53/18 144599 592714204} Qd7 {0.00/29 109707 126393915} 26. Re2 {0.52/19
319713 735593873} Qd6 {-0.18/32 106105 110201123} 27. b4 {0.45/19 319055
862263539} a6 {-0.18/27 102613 94765803} 28. Rb1 {0.45/17 157502 235158673} b5
{-0.33/26 99209 99099049} 29. axb5 {0.23/18 229242 438254696} axb5 {-0.33/27
95917 98607711} 30. Qa2 $1 {0.60/17 58609 384849189. next c5 is good for White.
} bxc4 {-0.33/27 92708 88236715} 31. dxc4 {0.21/17 294948 1029088535. Strongly
threatening c5.} Rd8 {-0.33/26 89707 74919363} 32. Qb3 {0.07/18 171045
380097552} Rb7 {-0.33/22 86716 94878704} 33. Kg2 {-0.45/31 154943 439920921} (
33. Ree1 $11 {0.00/31}) 33... Rc8 {-0.39/20 83911 67889999} 34. Rc2 {0.00/18
208914 557713546} Rd7 {-0.30/21 81114 68056877} 35. Qf3 {-0.05/18 370927
794668441} f5 {-0.51/23 78418 53113076} 36. Qd3 {-0.24/18 205491 444252154} Kg8
{-0.45/23 75818 71456745} 37. h3 {-0.05/17 186506 442166473} Ra8 {-0.75/21
73315 54261618} 38. Kh2 {-1.27/23 144722 312021263} ({Better is} 38. Rcb2 $15 {
-0.55/27}) 38... Rda7 {-0.10/26 70915 57457912} (38... Rf7 $17 {-1.27/23} 39.
Rcb2 Rfa7) 39. Rb3 $2 {-4.91/24 473725 1302527395. White does not recover from
this.} (39. b5 $1 $11 {-0.10/26}) 39... Ra1 $19 {-1.23/21 68616 41269414} 40.
Qd2 {-0.92/18 141487 355528099} Qf6 {-1.92/21 66409 42472844} 41. b5 {-1.10/17
67 158468974} e4 {-2.16/25 456308 166328255} 42. b6 {-2.14/21 134108 1150774810
} e3 {-5.01/24 289010 193270583} 43. Qg2 {-2.16/20 61 724767716} f4 {-5.28/23
279409 93005855. ...f3 is the strong threat.} 44. b7 {-4.44/20 231266
1039097504} Rb8 {-5.91/24 270111 240217722. aiming for ...f3.} 45. Qf3 {
-4.78/20 107300 845714013} Qf5 {-6.18/23 261115 169514646} 46. Rcb2 {-4.54/20
40 678103234} d3 {-7.89/22 252524 200098911} 47. Rb1 {-5.72/20 32590 639289879}
Rxb1 {-12.51/22 317829 341418001} 48. Rxb1 {-6.32/16 23342 61972793} e2 {
-12.60/22 233525 164409421} 49. gxf4 {-9.69/20 232636 604636548} Rxb7 $3 {
-14.16/23 [#] 225807 166147447. Deflection, Promotion} 50. Rxb7 {-5.65/22
Endgame 676430 630141146} (50. Rh1 Re7) 50... e1=Q {-14.61/16 327412 35887805.
Black mates.} 51. Rb8+ {-11.89/18 134082 432762298} Kg7 {-22.29/16 311118
42028160} 52. Rb2 {-12.64/21 24 919224087} d2 {-13.83/13 295619 13395012} 53.
Qc3+ {-16.39/23 41 1058504920} Qf6 {M+39/13 67985 52253859} 54. Qxd2 {-12.96/
20 244846 1062835892} Qxd2+ {M+38/10 1380 293919} 55. Rxd2 {-13.06/17 15482
77544623} Qxf4+ {[#] M+37/10 253 137999. Double Attack} 56. Kg2 {-20.40/19
303014 1352169080} Qxd2+ {M+38/10 2365 1154877} 57. Kg3 {-22.72/19 346947
1379513881} g5 {M+16/10 4323 4257866} 58. d6 {-327.45/19 414020 1716934798} g4
{M+11/10 6376 9143362} 59. hxg4 {-14.46/19 192854 858704255} Qxd6+ {M+10/10
1571 626710} 60. Kf3 {-17.77/19 356887 1732545336} Qd3+ {M+9/10 638 480396} 61.
Kf4 {-327.37/19 249245 1205677681} Qxc4+ {M+8/10 279 94160} 62. Kf3 {-327.35/
18 94689 481214270} Kg6 {10 M+7/10 302} 63. Ke3 {-327.39/15 9410 46580286} Kg5
{10 M+6/10 285} 64. Kd2 {-327.51/13 1240 5767535} Kxg4 {10 M+5/10 302} 65. Ke3
{-327.53/12 146 622320} Kg3 {10 M+4/10 285} 66. Kd2 {-17.28/7 25 21651} Kf3 {
10 M+3/10 325} 67. Kd1 {-327.65/3 37 1193} Ke3 {10 M+2/10 310} 68. Ke1 {
-327.67/2 17 328} Qc1# {10 mate M+1/10 292. Weighted Error Value: White=0.66/
Black=0.12} 0-1
[/pgn]
PS: If this is a typo please let me know so I can upgrade the version number. The release notes do show 2.9 on the webpage.
Thanks
UPDATE: I modded the 2.8 link and changed it to 2.9 and the Download came down. So I will now upgrade. http://technochess.free.fr/Archive/Knightx29_64.zip
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
__________________________________________________________________
Ted Summers
-
Rebel
- Posts: 7017
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: KnightX 2.9 Winboard engine released
Put your engine as new entry for the Gambit Rating List.
Games : 1200
Elo pool : 2696
You can follow the progress on - http://rebel13.nl/b/grl.htm
Page refreshes itself every 30 seconds.
Games : 1200
Elo pool : 2696
You can follow the progress on - http://rebel13.nl/b/grl.htm
Page refreshes itself every 30 seconds.
90% of coding is debugging, the other 10% is writing bugs.
-
Rebel
- Posts: 7017
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: KnightX 2.9 Winboard engine released
KnightX 2.8 - 2648
Next : KnightX 2.9
Same URL.
Code: Select all
Gambit Rating List
Running : gauntlet KnightX 2.8
Time Control : Time control : 40/120
Games : 1200
Results from file gauntlet-knightx.pgn:
No. Name Win Draw Loss Unf. Score Games %
----------------------------------------------------------
1 Knightx 2.8 +388 =239 -573 *0 507.5 1200 42.3%
2 Zahak 5.0 +114 =35 -51 *0 131.5 200 65.8%
3 k2 099 +100 =43 -57 *0 121.5 200 60.8%
4 Dumb 1.8 +98 =44 -58 *0 120.0 200 60.0%
5 Fruit 2.1 +94 =41 -65 *0 114.5 200 57.2%
6 Supernova 2.3 +89 =41 -70 *0 109.5 200 54.8%
7 Monolith 0.4 +78 =35 -87 *0 95.5 200 47.8%
Total Games: 1200
White Wins: 465 (38.8%)
Black Wins: 496 (41.3%)
Draws: 239 (19.9%)
Unfinished: 0 (0.0%)
Estimated ratings for this elo 2696 pool
# PLAYER : RATING POINTS PLAYED (%)
1 Zahak 5.0 : 2763.0 131.5 200 66
2 k2 099 : 2725.3 121.5 200 61
3 Dumb 1.8 : 2719.8 120.0 200 60
4 Fruit 2.1 : 2699.9 114.5 200 57
5 Supernova 2.3 : 2682.2 109.5 200 55
6 Knightx 2.8 : 2648.8 507.5 1200 42
7 Monolith 0.4 : 2633.0 95.5 200 48Same URL.
90% of coding is debugging, the other 10% is writing bugs.
-
Rebel
- Posts: 7017
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: KnightX 2.9 Winboard engine released
Sadly I didn't notice a 50 elo improvement, instead a regression.
2.8 - 2648
2.9 - 2623
Code: Select all
Gambit Rating List
Running : gauntlet KnightX 2.9
Time Control : Time control : 40/120
Games : 1200
Results from file gauntlet-knightx-29.pgn:
No. Name Win Draw Loss Unf. Score Games %
----------------------------------------------------------
1 Knightx 2.9 +333 =253 -614 *0 459.5 1200 38.3%
2 Zahak 5.0 +121 =30 -49 *0 136.0 200 68.0%
3 k2 099 +107 =42 -51 *0 128.0 200 64.0%
4 Dumb 1.8 +109 =36 -55 *0 127.0 200 63.5%
5 Supernova 2.3 +102 =48 -50 *0 126.0 200 63.0%
6 Fruit 2.1 +99 =42 -59 *0 120.0 200 60.0%
7 Monolith 0.4 +76 =55 -69 *0 103.5 200 51.8%
Total Games: 1200
White Wins: 466 (38.8%)
Black Wins: 481 (40.1%)
Draws: 253 (21.1%)
Unfinished: 0 (0.0%)
Estimated ratings for this elo 2696 pool
# PLAYER : RATING POINTS PLAYED (%)
1 Zahak 5.0 : 2755.7 136.0 200 68
2 k2 099 : 2724.5 128.0 200 64
3 Dumb 1.8 : 2720.7 127.0 200 64
4 Supernova 2.3 : 2716.9 126.0 200 63
5 Fruit 2.1 : 2694.7 120.0 200 60
6 Monolith 0.4 : 2635.9 103.5 200 52
7 Knightx 2.9 : 2623.6 459.5 1200 382.9 - 2623
90% of coding is debugging, the other 10% is writing bugs.
-
chrjly
- Posts: 37
- Joined: Wed Feb 03, 2021 10:45 pm
- Full name: Christophe Jolly
Re: KnightX 2.9 Winboard engine released
You are right. There were some regressions in the search. It could loose a draw game or draw a winning game sometimes. It is fixed in v2.91, anyways I hope so.