Computer draughts tournament, Culemborg (2011.09.18)

[Ed Gilbert]

Preliminary announcement for the open Dutch championship rapid Polish Draughts for computerprograms 2011.

Day of the tournament: Sunday the 18th of September 2011.

Location: PW Consulting, Stationssingel 25 Culemborg, The Netherlands.

Time: 09:30 Installing the hardware, 10.00 Start first round.


Rules will be the same as during the tournaments of recent date (last year we played 7 rounds with playing-time 20 minutes per round per participant; after 75 moves arbitration if necessary).
Jaap Bus will again be the referee (well, this year only partially).


If you want to participate, please report by answering this e-mail.
Please report your participation before the 11th of September.

frgr

Leo Nagels

P.S. If you have participated in the voting a month ago and reported non-participation for this date, you won't need to reply anymore; if you have participated in the voting a month ago and reported participation (certain or probable), please confirm your participation once more.

[Ed Gilbert]

Heerlen, 9 september 2011


To whom it may concern:



Second announcement for the open Dutch championship rapid Polish Draughts for computerprograms 2011.

Day of the tournament: Sunday the 18th of September 2011.

Location: PW Consulting, Stationssingel 25 Culemborg, The Netherlands.

Time: 09:30 Installing the hardware, 10.00 Start first round.

List of participants:


Boase Krijn van Drie
Cerberus Leo Nagels
Dam 2.2 Harm Jetten
DIOS 2000 Chris Jurriëns
Dream Said Koudache
GWD Klaas Bor / Gijsbert Wiessenekker
Maximus Jan-Jaap van Horssen
Sjende Blyn Jelle Wiersma
TD King Ton Tillemans
Tornado Frank Mesander

As you see, we have a debutant Krijn van Drie with Boase, watch out for him, he is from Culemborg, so he has the home advantage!

Rules will be the same as during the tournaments of recent date: 9 rounds with playing-time 20 minutes per round per participant; after 75 moves arbitration if necessary.

See you in Culemborg,

Leo Nagels

[Ed Gilbert]

Congrats to Jan-Jaap van Horssen for winning the tournament with Maximus.

1. Maximus 14 points
2. DIOS
Sjende Blyn 13 points
4. TD King 12 points
5. Tornado 11 points
6. GWD 9 points
7. Dam 2.2 7 points
8. Cerberus
Dream 5 points
10. Boase 1 point

[BertTuyt]

Jan-Jaap, great result !!!
In Dutch wat goed is komt snel

Could you share some details related to the setup used , such as Hardware/Software, 64-bit OS, bitboard, nodes/second, parallel search, endgame DB, and (if any) some results against some existing programs..

Bert

[jj]

Jan-Jaap, great result !!!
In Dutch wat goed is komt snel

Could you share some details related to the setup used , such as Hardware/Software, 64-bit OS, bitboard, nodes/second, parallel search, endgame DB, and (if any) some results against some existing programs..

Bert

Bert, Ed, and the people who emailed me personally,

Thank you. Here are some details about the setup used.

The computer I used during the tournament (actually my brother's home cinema computer) is an Intel Quad Core i7-2600K with 16Gb RAM, running Windows 7 Ultimate 64-bit. Maximus is written in Java and runs on a 64-bit Java virtual machine. It uses bitboards and a parallel search (YBWC), and 2-6 pieces endgame databases, these were generated using the database generator by Michel Grimminck and Harm Jetten. During the tournament the average speed was 15.5 Mnodes/sec.

I haven't played any matches against some of the existing programs yet with this Maximus version.

Bert, I'm also interested to know what the Damage setup etc. was in last year's tournament?

Jan-Jaap

[BertTuyt]

Jan-Jaap, last year I used (and still use) a i7-940 (2.93 Ghz), 12 Mbyte Memory 7p DB, bitboards, parallel search (YBWC, much "learned" from Glaurung).
Speed (4 cores) around 8 - 10 MNodes/Sec.
To increase the effectiveness of the endgame DB I stored the (around) 30 GByte on a SSD.
So far the only match lost is against Kingsrow (still to my opnion the program to beat ).
Matches against Flits and Truus are a win for Damage.

Bert

[BertTuyt]

Forgot to mention Windows 7 64Bit...

Bert

[Piet Bouma]

Is there anywhere (or could anyone supply) on the Internet the cross-table, results and games in .pdn of the Tournament?
So I could place the tournament in Toernooibase/Tournamentbase?

[Ed Gilbert]

Is there anywhere (or could anyone supply) on the Internet the cross-table, results and games in .pdn of the Tournament?
So I could place the tournament in Toernooibase/Tournamentbase?

I now have this info. If you still need it send me an email and I will send you the files.

-- Ed

[Piet Bouma]

Is there anywhere (or could anyone supply) on the Internet the cross-table, results and games in .pdn of the Tournament?
So I could place the tournament in Toernooibase/Tournamentbase?

I now have this info. If you still need it send me an email and I will send you the files.

-- Ed

TD King has send me also the files.
Matrix and games on applet in Toernooibase/Tournamentbase

[jj]

Could you share [...] some results against some existing programs..

Hi all,

Over the past two weeks I ran DXP matches against Dam2.2 (no Harm intended), Truus, Flits and Kingsrow with the Maximus version that played the Culemborg tournament. In exactly the same configuration, just with different time controls. During the matches each program ran on a separate computer, so it could use all the resources it needed. Maximus had a little advantage because it was running on a somewhat faster machine. And of course Dam, Truus and Flits do not have a parallel search.

The results:

Code: Select all

Maximus 0.88 vs Dam2.2 v.7       72 wins,  1 loss  ,  85 draws (73%)
Maximus 0.88 vs Truus v.Win 1.0  27 wins, 14 losses, 117 draws (54%)
Maximus 0.88 vs Flits 3.02       11 wins, 20 losses, 127 draws (47%)
Maximus 0.88 vs Kingsrow 1.51     2 wins, 38 losses, 118 draws (39%)

Hardware used and program settings:

Spoiler:

Maximus settings:
--------------------
Hardware: Intel Quad Core i7-2600K 3.4 GHz 16GB RAM
Time control: 5 seconds fixed per move
Hashtable size: 32 M positions
Egdb: 2-6 pieces
Opening book: best moves
Pondering: enabled (7 threads)
Search (hyper)threads: 8

Dam2.2 settings:
-------------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Pondering: no (is disabled during DXP games)
Geheugen: 'Enorm veel' (24M)
Egdb: 2-6 pieces

Truus settings:
-----------------
Hardware: Intel Dual Core E8400 3.0 GHz 4GB RAM
Time control: 75 moves in 5 minutes
Pondering: yes
Partijen database ruimte (MB): 4 (the default; I should raise this to 35 MB, the size of trubase.tp)
Stellingen ruimte (16K): 10000
Egdb: 2-6 pieces

Flits settings:
---------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Pondering: yes
Transpositietabel (MB): 256
Egdb: 2-6 pieces (all enabled)

Kingsrow settings:
--------------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Hashtable size: 512mb
Egdb cache: 9000mb
Egdb subsets: 2-8 pieces + 9 pieces men only
Opening book: best moves
Pondering: enabled
Search (hyper)threads: 8

I should probably repeat these matches a couple of times to get a more reliable picture, but my brother won't let me "burn" his video/game computer too much :-). Maximus has improved since the previous version, but Flits is still an angstgegner! Kingsrow seems the ultimate program to beat...

Bert, what are the Damage scores against these programs?

Jan-Jaap

[Rein Halbersma]

Could you share [...] some results against some existing programs..

Hi all,

Over the past two weeks I ran DXP matches against Dam2.2 (no Harm intended), Truus, Flits and Kingsrow with the Maximus version that played the Culemborg tournament. In exactly the same configuration, just with different time controls. During the matches each program ran on a separate computer, so it could use all the resources it needed. Maximus had a little advantage because it was running on a somewhat faster machine. And of course Dam, Truus and Flits do not have a parallel search.

The results:

Code: Select all

Maximus 0.88 vs Dam2.2 v.7       72 wins,  1 loss  ,  85 draws (73%)
Maximus 0.88 vs Truus v.Win 1.0  27 wins, 14 losses, 117 draws (54%)
Maximus 0.88 vs Flits 3.02       11 wins, 20 losses, 127 draws (47%)
Maximus 0.88 vs Kingsrow 1.51     2 wins, 38 losses, 118 draws (39%)

Hardware used and program settings:

Spoiler:

Maximus settings:
--------------------
Hardware: Intel Quad Core i7-2600K 3.4 GHz 16GB RAM
Time control: 5 seconds fixed per move
Hashtable size: 32 M positions
Egdb: 2-6 pieces
Opening book: best moves
Pondering: enabled (7 threads)
Search (hyper)threads: 8

Dam2.2 settings:
-------------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Pondering: no (is disabled during DXP games)
Geheugen: 'Enorm veel' (24M)
Egdb: 2-6 pieces

Truus settings:
-----------------
Hardware: Intel Dual Core E8400 3.0 GHz 4GB RAM
Time control: 75 moves in 5 minutes
Pondering: yes
Partijen database ruimte (MB): 4 (the default; I should raise this to 35 MB, the size of trubase.tp)
Stellingen ruimte (16K): 10000
Egdb: 2-6 pieces

Flits settings:
---------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Pondering: yes
Transpositietabel (MB): 256
Egdb: 2-6 pieces (all enabled)

Kingsrow settings:
--------------------
Hardware: Intel Quad Core i7-920 2.79 GHz 12GB RAM
Time control: 75 moves in 5 minutes
Hashtable size: 512mb
Egdb cache: 9000mb
Egdb subsets: 2-8 pieces + 9 pieces men only
Opening book: best moves
Pondering: enabled
Search (hyper)threads: 8

I should probably repeat these matches a couple of times to get a more reliable picture, but my brother won't let me "burn" his video/game computer too much . Maximus has improved since the previous version, but Flits is still an angstgegner! Kingsrow seems the ultimate program to beat...

Bert, what are the Damage scores against these programs?

Jan-Jaap

Hi Jan-Jaap,

Those are very nice match results! Just in between Truus & Flits rating-wise. Could you post the 2 wins against Kingsrow?

Rein

[jj]

Hi Jan-Jaap,

Those are very nice match results! Just in between Truus & Flits rating-wise. Could you post the 2 wins against Kingsrow?

Rein

Hi Rein,

Thank you. Did you already play DXP matches with <Mistral>?
Below are the 2 wins against Kingsrow.

Jan-Jaap

[Event "Game 113, opening 33-28 20-25"]
[Date "2011.09.29"]
[White "Kingsrow 1.51"]
[Black "Maximus 0.88a"]
[Result "0-1"]
[FEN "W:W28,31,32,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50:B1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,25."]
1. 32-27 19-23 2. 28x19 14x23 3. 39-33 10-14 4. 44-39 14-20 5. 37-32 5-10 6.
41-37 13-19 7. 50-44 10-14 8. 31-26 8-13 9. 46-41 2-8 10. 36-31 17-21 11. 26x17
12x21 12. 33-28 21-26 13. 39-33 16-21 14. 27x16 7-12 15. 16x7 6-11 16. 7x16
18-22 17. 28x17 12x21 18. 16x27 23-29 19. 33x24 19x50 20. 41-36 50-11 21. 40-34
11-2 22. 38-33 14-19 23. 42-38 1-7 24. 34-29 13-18 25. 32-28 19-23 26. 28x19
20-24 27. 19x30 25x23 28. 38-32 7-12 29. 43-38 9-13 30. 49-43 15-20 31. 45-40
4-10 32. 47-42 10-15 33. 43-39 12-17 34. 27-21 20-25 35. 21x12 8x17 36. 33-28
2-8 37. 28x19 13x24 38. 40-34 18-22 39. 39-33 22-28 40. 33x11 24-30 41. 35x24
8x6 42. 31-27 6-1 43. 27-22 25-30 44. 38-33 1-7 45. 22-17 3-8 46. 33-29 7x40 47.
17-11 30-35 48. 32-28 40-1 49. 11-6 35-40 50. 42-38 8-12 51. 28-23 40-44 52.
23-19 44-49 53. 48-43 49-35 54. 19-14 35-13 55. 38-32 12-18 56. 32-28 13-8 57.
28-23 18x29 58. 14-9 29-34 59. 43-38 34-40 60. 9-4 40-45 61. 4-27 0-1

[Event "Game 135, opening 34-29 18-22"]
[Date "2011.09.29"]
[White "Kingsrow 1.51"]
[Black "Maximus 0.88a"]
[Result "0-1"]
[FEN "W:W29,31,32,33,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50:B1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,19,20,22."]
1. 40-34 12-18 2. 44-40 7-12 3. 31-27 22x31 4. 37x26 19-23 5. 35-30 20-25 6.
30-24 14-20 7. 41-37 10-14 8. 40-35 17-22 9. 46-41 25-30 10. 34x25 23x34 11.
39x30 20x29 12. 33x24 22-28 13. 32x23 18x20 14. 37-32 5-10 15. 50-44 20-24 16.
30x19 14x23 17. 41-37 10-14 18. 35-30 14-20 19. 25x14 9x20 20. 38-33 13-19 21.
30-25 20-24 22. 43-38 8-13 23. 36-31 1-7 24. 44-40 4-9 25. 48-43 12-18 26. 31-27
2-8 27. 47-41 15-20 28. 25x14 19x10 29. 41-36 10-14 30. 43-39 14-19 31. 49-44
9-14 32. 33-28 14-20 33. 37-31 7-12 34. 42-37 20-25 35. 40-35 24-29 36. 39-33
29-34 37. 33-29 34-39 38. 44x33 23x34 39. 35-30 34-39 40. 33x44 25x34 41. 38-33
12-17 42. 44-40 34-39 43. 33x44 17-21 44. 26x17 11x33 45. 40-35 19-24 46. 45-40
24-29 47. 44-39 33x44 48. 40x49 29-34 49. 32-28 8-12 50. 49-43 3-9 51. 28-22
9-14 52. 31-26 18-23 53. 26-21 23-29 54. 22-17 14-19 55. 17x8 13x2 56. 21-17
29-33 57. 35-30 34x25 58. 17-12 25-30 59. 36-31 0-1

[Rein Halbersma]

Hi Jan-Jaap,

Those are very nice match results! Just in between Truus & Flits rating-wise. Could you post the 2 wins against Kingsrow?

Rein

Hi Rein,

Thank you. Did you already play DXP matches with <Mistral>?
Below are the 2 wins against Kingsrow.

Jan-Jaap

No, other (both work and private) projects have kept me from completing my engine, and it is not ready to play other programs yet.

Nice wins BTW! Especially surprising that Kingsrow allowed the opening combination in game 1, as it usual is not afraid to take such expensive king combinations itself. You seem to have some good king placement code, or was that placement of the black king behind friendly lines more or less search-guided?

Rein

[jj]

No, other (both work and private) projects have kept me from completing my engine, and it is not ready to play other programs yet.

Nice wins BTW! Especially surprising that Kingsrow allowed the opening combination in game 1, as it usual is not afraid to take such expensive king combinations itself. You seem to have some good king placement code, or was that placement of the black king behind friendly lines more or less search-guided?

Rein

Thanks. I have to admit that the king placement was search-guided...

Jan-Jaap