Unofficial World Championship Of Computer Programs In International Draughts 2021

[Krzysztof Grzelak]

You must write to the author of the program.

[clp]

There is obviously no compelling reason to tie-break.

A correct sequence without tie-breaking is:

Place No. Program     Pts
1-1    4  Damage NNUE 54
2-6    1  Ares NNUE   50
2-6    6  Dragon      50
2-6   10  Kingsrow    50
2-6   14  Scan        50
2-6   15  Sjende Blyn 50
7-7   11  Maximus     48

[clp]

You must write to the author of the program.

The program is OK. The usage of the program is wrong. Consult an expert.

[EricvanDusseldorp]

I completely agree with Hendrik and Kees.
There is a difference in the functioning of the Draughts Arbiter program and the tournament rules.
And in those cases the rules of the tournament take precedence.
The regulations are leading, the program is next. Not the other way around.

Całkowicie zgadzam się z Hendrik i Kees.
Istnieje różnica w funkcjonowaniu programu Draughts Arbiter i zasadach turnieju.
I w takich przypadkach pierwszeństwo mają reguły turnieju.
Przepisy prowadzą, program następny. Nie na odwrót.

[Krzysztof Grzelak]

I completely agree with Hendrik and Kees.
There is a difference in the functioning of the Draughts Arbiter program and the tournament rules.
And in those cases the rules of the tournament take precedence.
The regulations are leading, the program is next. Not the other way around.

Całkowicie zgadzam się z Hendrik i Kees.
Istnieje różnica w funkcjonowaniu programu Draughts Arbiter i zasadach turnieju.
I w takich przypadkach pierwszeństwo mają reguły turnieju.
Przepisy prowadzą, program następny. Nie na odwrót.

What's the difference. Dragon won 17 games and Sjende Blyn 16 games. How can Sjende Blyn be before Dragon.

[EricvanDusseldorp]

I completely agree with Hendrik and Kees.
There is a difference in the functioning of the Draughts Arbiter program and the tournament rules.
And in those cases the rules of the tournament take precedence.
The regulations are leading, the program is next. Not the other way around.

Całkowicie zgadzam się z Hendrik i Kees.
Istnieje różnica w funkcjonowaniu programu Draughts Arbiter i zasadach turnieju.
I w takich przypadkach pierwszeństwo mają reguły turnieju.
Przepisy prowadzą, program następny. Nie na odwrót.

What's the difference. Dragon won 17 games and Sjende Blyn 16 games. How can Sjende Blyn be before Dragon.

Because Dragon lost one game and Sjende Blyn didn't.
So equal points for both.
And then SB is decisive, according to your own rules.

[Krzysztof Grzelak]

Because Dragon lost one game and Sjende Blyn didn't.
So equal points for both.
And then SB is decisive, according to your own rules.

This has nothing to do with it. The most important thing is winning games, then draws and finally losing games.This is the case at every single tournament.

[EricvanDusseldorp]

I rest my case.

[Sidiki]

Hi all,

I see about what your aren't agree with each other:

One part write about strenght difference, another part of won games linked to won points.

It's like a program that have an elo of 3000 points that draw against another one with 2200 elo must lose some points and the weaker must win points.
In this logic, if among some programs rated at the same strenght one of them lose a game and the other draw all they game, the first one must be underclassed. Damage 16.1 NNUE and Kingsrow 1.63 among the strongest programs drawed against Dream 3x3, i suppose in this case that they must to lose point!?

The other side is in another logic, if 6 engines participate to a tournament and one lose 2 games / 6 and win the 4 games left .
If the 5 others engines draw all they games between themselves excepted each game won against wich that lost the 2 games, the 1rst will have 8 points, 2nds 6 points and the 4th.

It will be simple to not take into account the strenght difference but the number of points won.

Friendly, Sidiki

[Yves]

Hello everyone.

Thanks to Krzysztof Grzelak and the software programmers.

By respecting the calculation of SB points of DraughtsArbitrePro,
the real score is based on games won and lost.

points won no lost SB score
Damage 54 20 14 0 1589 1
SjendeBlyn 50 16 18 0 1448 2
Kingsrow 50 16 18 0 1444 3
Scan 50 16 18 0 1437 4
Ares 50 16 18 0 1433 5
Dragon 50 17 16 1 1398 6
Maximus 48 14 20 0 1417 7

[jj]

Of course there was also the first game Dragon-Ares in round 23, which would have been a win for Dragon if continued. That has more impact on the GC than SB or number of wins.

Code: Select all

1.   Damage NNUE                 54
2.   Dragon                      51
3-5. Kingsrow, Scan, Sjende Blyn 50
6.   Ares NNUE                   49
7.   Maximus                     48

What do people think, was replaying that game later a fair solution? In ICGA tournaments you are allowed to recover from a crash but here there is no rule for this. What should be the rule?

Game Tdking - Ares repeated due to technical issues.

It is good that you write something on the forum if something happened, but this does not give any information. What technical issues? Did one of the programs fail or was it something else?

What happens for instance if you accidentally make the wrong move on the DGT board? Does the software allow you to recover from that or do you have to start a new game? The point is that the game operator should not have any influence on the outcome of the games, which now seems not to be guaranteed.

I think that with the increasing importance of Krzysztof's tournament (as opposed to the Computer Olympiad) the rules need to be clear and fair on these points and agreed upon by the programmers. And of course carefully formulated on the web site and followed.

[Krzysztof Grzelak]

Of course there was also the first game Dragon-Ares in round 23, which would have been a win for Dragon if continued. That has more impact on the GC than SB or number of wins.

Code: Select all

1.   Damage NNUE                 54
2.   Dragon                      51
3-5. Kingsrow, Scan, Sjende Blyn 50
6.   Ares NNUE                   49
7.   Maximus                     48

What do people think, was replaying that game later a fair solution? In ICGA tournaments you are allowed to recover from a crash but here there is no rule for this. What should be the rule?

Game Tdking - Ares repeated due to technical issues.

It is good that you write something on the forum if something happened, but this does not give any information. What technical issues? Did one of the programs fail or was it something else?

What happens for instance if you accidentally make the wrong move on the DGT board? Does the software allow you to recover from that or do you have to start a new game? The point is that the game operator should not have any influence on the outcome of the games, which now seems not to be guaranteed.

I think that with the increasing importance of Krzysztof's tournament (as opposed to the Computer Olympiad) the rules need to be clear and fair on these points and agreed upon by the programmers. And of course carefully formulated on the web site and followed.

I think very well that I did it, at least I was fair. I will write why in a moment. If this should happen in a real tournament, the Ares program should have a winning game. Remember one thing, a winning position does not mean a winning game. I will add from myself that I would do it with every engine in the tournament. The technical problem was for another reason. What will happen if I make a mistake on the DGT board. You have to repeat the whole game all over again.

[jj]

If this should happen in a real tournament, the Ares program should have a winning game.

Not in an ICGA tournament, as I said.

The technical problem was for another reason.

Which was?

What will happen if I make a mistake on the DGT board. You have to repeat the whole game all over again.

I wrote software for the DGT board that allows you to transmit games on the internet and that allows you to take back mistakes. Let me know if you are interested in this for next year's tournament.

[clp]

What do people think, was replaying that game later a fair solution? In ICGA tournaments you are allowed to recover from a crash but here there is no rule for this. What should be the rule?

What to do in case of a crash should be in the tournament regulations ideally. If not the operator handles it to its own discretion (as Krzysztof did).

Why not continue the game from the last move ? Starting the game over again is unsatisfactory in my opinion.

[Krzysztof Grzelak]

What to do in case of a crash should be in the tournament regulations ideally. If not the operator handles it to its own discretion (as Krzysztof did).

Why not continue the game from the last move ? Starting the game over again is unsatisfactory in my opinion.

DGT board does not allow this.