Internet engine matches

Discussion about development of draughts in the time of computer and Internet.
Post Reply
Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Thu Sep 07, 2017 09:49

Match KINGSROW - DRAGON (2-move ballots)

Kingsrow 1.58 vs. Dragon 4.6 8 wins, 2 losses, 148 draws, 0 unknowns

Kingsrow

Opening Book = Best Moves
HashTable Size = 64 MB
DB cache Size = 12000 MB
The base ends = 6 Pieces
Threads = 4
Time = 3 Min / 100 Moves

Dragon

Opening Book = Tournament Book
HashTable Size = 64 MB
The base ends = 6 Pieces
Avoid draws - Not
Number of cores = 4
Time = 3 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
Attachments
dxpgames.pdn
(164.64 KiB) Downloaded 128 times

BertTuyt
Posts: 1303
Joined: Wed Sep 01, 2004 19:42

Re: Internet engine matches

Post by BertTuyt » Sat Sep 09, 2017 12:34

Krzysztof, thanks for sharing and doing these matches.

It really seems not easy to defeat Scan :D

Bert

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Sat Sep 09, 2017 20:13

Fabien spoke very well about the program Dragon. I have played several games with different programs and saw and evaluated new possibilities of the program Dragon.

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Sat Sep 23, 2017 22:02

Match KINGSROW - SCAN (2-move ballots)

Kingsrow 1.59 vs. Scan 3.0 0 wins, 6 losses, 152 draws, 0 unknowns

Kingsrow 1.59

Opening Book = Best Moves
HashTable Size = 32 MB
DB cache Size = 10000 MB
The base ends = 8 Pieces
Threads = 4
Time = 1 Min / 100 Moves

Scan 3.0

Variant = normal
Book = true
Book-ply = 4
Book-margin = 4
Threads = 4
Tt-size = 24
Bb-size = 6
Time = 1 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
Attachments
dxpgames.pdn
(153.97 KiB) Downloaded 136 times

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Mon Sep 25, 2017 18:11

Match KINGSROW - SCAN (2-move ballots)

Kingsrow 1.59 vs. Scan 3.0 0 wins, 5 losses, 153draws, 0 unknowns

Kingsrow 1.59

Opening Book = Best Moves
HashTable Size = 64 MB
DB cache Size = 10000 MB
The base ends = 8
Threads = 4
Time = 3 Min / 100 Moves

Scan 3.0

Variant = normal
Book = true
Book-ply = 4
Book-margin = 4
Threads = 4
Tt-size = 24
Bb-size = 6
Time = 3 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
Attachments
dxpgames.pdn
(150.25 KiB) Downloaded 126 times

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Tue Sep 26, 2017 12:27

Match KINGSROW - SCAN (2-move ballots)

Kingsrow 1.59 vs. Scan 3.0 0 wins, 2 losses, 156 draws, 0 unknowns

Kingsrow 1.59

Opening Book = Best Moves
HashTable Size = 128 MB
DB cache Size = 10000 MB
The base ends = 8
Threads = 4
Time = 5 Min / 100 Moves

Scan 3.0

Variant = normal
Book = true
Book-ply = 4
Book-margin = 4
Threads = 4
Tt-size = 24
Bb-size = 6
Time = 5 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
Attachments
dxpgames.pdn
(151.11 KiB) Downloaded 127 times

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Internet engine matches

Post by TAILLE » Tue Sep 26, 2017 18:30

Krzysztof Grzelak wrote:Match KINGSROW - SCAN (2-move ballots)

Kingsrow 1.59 vs. Scan 3.0 0 wins, 2 losses, 156 draws, 0 unknowns

Kingsrow 1.59

Opening Book = Best Moves
HashTable Size = 128 MB
DB cache Size = 10000 MB
The base ends = 8
Threads = 4
Time = 5 Min / 100 Moves

Scan 3.0

Variant = normal
Book = true
Book-ply = 4
Book-margin = 4
Threads = 4
Tt-size = 24
Bb-size = 6
Time = 5 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
As far as I am concerned the results of these Scan - Kingsrow matches are not good news for international 10x10 draughts game because I see the number of draws becoming nearer and nearer from 100%!
Difficult in this context to keep the motivation.
Gérard

BertTuyt
Posts: 1303
Joined: Wed Sep 01, 2004 19:42

Re: Internet engine matches

Post by BertTuyt » Tue Sep 26, 2017 19:29

First of all thanks for organizing the matches.
I also tend to agree with Gerard, this is close at 100% draw level.
And not yet under normal tournament conditions, or hyper Hardware.
Think in this case the ELO differences was 4 points.
But the questions remains, are we close to the draw black hole, or are there still many ELO points to earn moving towards perfection.

Bert

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Tue Sep 26, 2017 21:21

BertTuyt wrote:First of all thanks for organizing the matches.
I also tend to agree with Gerard, this is close at 100% draw level.
And not yet under normal tournament conditions, or hyper Hardware.
Think in this case the ELO differences was 4 points.
But the questions remains, are we close to the draw black hole, or are there still many ELO points to earn moving towards perfection.

Bert
I have the impression that you forgot what is most important in the draughts.

BertTuyt
Posts: 1303
Joined: Wed Sep 01, 2004 19:42

Re: Internet engine matches

Post by BertTuyt » Tue Sep 26, 2017 21:55

which is??

Bert

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Thu Sep 28, 2017 07:30

BertTuyt wrote:which is??

Bert
Winning. Elo in draughts does not play.

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Thu Sep 28, 2017 07:44

Match KINGSROW - SCAN (2-move ballots)

Kingsrow 1.59 vs. Scan 3.0 0 wins, 2 losses, 156 draws, 0 unknowns

Kingsrow 1.59

Opening Book = Best Moves
HashTable Size = 256 MB
DB cache Size = 10000 MB
The base ends = 8
Threads = 4
Time = 10 Min / 100 Moves

Scan 3.0

Variant = normal
Book = true
Book-ply = 4
Book-margin = 4
Threads = 4
Tt-size = 24
Bb-size = 6
Time = 10 Min / 100 Moves


Match played on a computer with the equipment.

Processor - Intel Core I7 2670QM, 2.2GHz
Hard disc - SSD Samsung 840 Evo 1 TB
Memory of frames - 16 GB DDR3 1333
System - Windows 7 Home Premium 64 bit Service Pack 1 PL
Attachments
dxpgames.pdn
(149.12 KiB) Downloaded 131 times

TAILLE
Posts: 968
Joined: Thu Apr 26, 2007 18:51
Location: FRANCE

Re: Internet engine matches

Post by TAILLE » Thu Sep 28, 2017 12:52

Krzysztof Grzelak wrote:
BertTuyt wrote:which is??

Bert
Winning. Elo in draughts does not play.
I fully agree with you Krzysztof , the most important point is winning!

10 or 20 years ago the best programs were not so strong and we observed a lot of winning games.
Unfortunately my feeling is that draughts is really a drawing game : during a high level game I think that almost all positions encountered are drawing positions in which it exists on average several drawing moves (the others being losing moves). The point is that the best programs are now almost always able to avoid the losing moves and, as a consequence, the number of draws becomes near from 100%.
In this sense, for the programmer I am, international draughts games, becomes less interesting.

For that reason, I was quite happy and motivated, to follow Bert work on the breakthrougth draughts. The 8x8 BT is now solved but we perfectly now that 10x10 BT cannot be solved by current computers. In addition, from the experience on the 8x8 BT, we know for sure that the game is very difficult and we know that it is extremely easy to transform a winning game into a losing one!
Because the draw do not exist the 10x10 BT, this game is for me a fascinating alternative to the international 10x10 game.
Gérard

Krzysztof Grzelak
Posts: 705
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Internet engine matches

Post by Krzysztof Grzelak » Thu Sep 28, 2017 13:45

TAILLE wrote: I fully agree with you Krzysztof , the most important point is winning!

10 or 20 years ago the best programs were not so strong and we observed a lot of winning games.
Unfortunately my feeling is that draughts is really a drawing game : during a high level game I think that almost all positions encountered are drawing positions in which it exists on average several drawing moves (the others being losing moves). The point is that the best programs are now almost always able to avoid the losing moves and, as a consequence, the number of draws becomes near from 100%.
In this sense, for the programmer I am, international draughts games, becomes less interesting.

For that reason, I was quite happy and motivated, to follow Bert work on the breakthrougth draughts. The 8x8 BT is now solved but we perfectly now that 10x10 BT cannot be solved by current computers. In addition, from the experience on the 8x8 BT, we know for sure that the game is very difficult and we know that it is extremely easy to transform a winning game into a losing one!
Because the draw do not exist the 10x10 BT, this game is for me a fascinating alternative to the international 10x10 game.
Write honestly and openly what I think.

As soon as Scan 2.0 and 3.0 were created, you Gerard and Bert are suspended and you are no longer trying to do a better program than the program Scan. What else would you write and would you otherwise write when the program was the best. As for Bert, it is Bert who writes a lot and does little, and in addition he is a liar. He promises to the public on the forum, and later withdraws from it.

Rein Halbersma
Posts: 1625
Joined: Wed Apr 14, 2004 16:04
Contact:

Re: Internet engine matches

Post by Rein Halbersma » Thu Sep 28, 2017 15:28

Krzysztof Grzelak wrote:
TAILLE wrote: I fully agree with you Krzysztof , the most important point is winning!

10 or 20 years ago the best programs were not so strong and we observed a lot of winning games.
Unfortunately my feeling is that draughts is really a drawing game : during a high level game I think that almost all positions encountered are drawing positions in which it exists on average several drawing moves (the others being losing moves). The point is that the best programs are now almost always able to avoid the losing moves and, as a consequence, the number of draws becomes near from 100%.
In this sense, for the programmer I am, international draughts games, becomes less interesting.

For that reason, I was quite happy and motivated, to follow Bert work on the breakthrougth draughts. The 8x8 BT is now solved but we perfectly now that 10x10 BT cannot be solved by current computers. In addition, from the experience on the 8x8 BT, we know for sure that the game is very difficult and we know that it is extremely easy to transform a winning game into a losing one!
Because the draw do not exist the 10x10 BT, this game is for me a fascinating alternative to the international 10x10 game.
Write honestly and openly what I think.

As soon as Scan 2.0 and 3.0 were created, you Gerard and Bert are suspended and you are no longer trying to do a better program than the program Scan. What else would you write and would you otherwise write when the program was the best. As for Bert, it is Bert who writes a lot and does little, and in addition he is a liar. He promises to the public on the forum, and later withdraws from it.
Krzysztof, please consider your tone and drop the accusations, please. All of us here are unpaid hobbyists trying to write our own programs. We have jobs and families that take priority over hobbies. If anyone makes announcements, they are intentions, not promises like a company would make to its customers. Scan, Kingsrow and Dragon have raised the bar to such heights that beating this level is very hard.

Post Reply