Computer draughts tournament announcement

[Ed Gilbert]

Heerlen, 2nd july 2009

To whom it may concern:



Preliminary announcement for the Mindbus Dutch open rapid draughts tournament for computers 2009.

Day of the tournament: Sunday the 13th of September 2009.

Location: Denksport Centrum Leiden (translation : "Mind Sport Centre Leiden")
Address: Robijnstraat 4, 2332 KE, Leiden (the French say Leyde), The Netherlands.
Tel.: 071 - 5 723 229

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

We play the tournament on the specified location by invitation of
draughts-club Leids Damgenootschap.
On februari 10 next year they will be 100 years of age, their first Centennial!
Mindbus is an ICT-company that will sponsor the tournament.
Maurits Meijer can be contacted on behalf of the draughts-club LDG,
E-mail: meijer20@gmail.com
As you may have noticed : this year the tournament is a few month earlier than the preceding years.
We hope you will appreciate this.
Rules will be the same as during the tournaments of recent date (last year
we played 9 rounds with playing-time 20 minutes per round per participant;
after 75 moves arbitration if necessary).
Jaap Bus will again be the referee.


If you want to participate, please report by answering this e-mail.
Please report your participation before the 25th of August.



Fr gr,



Leo Nagels

[TAILLE]

Hi,

Two questions :
Can somebody tell me how will be resolved a tie break ?
Are you in favor to play quicker games in order to play 2 games against each opponent ?

[Ed Gilbert]

Gerard,

I will probably not be at this tournament, but I am interested in your questions and improving the tournaments. I think if we want to see any changes in the rules we should decide exactly what we want and then propose them to Leo and the other programmers. In addition to your questions about tie break and playing more games, I have a few others:

- Under what conditions can the game clock be paused? Can a program be restarted if it appears to be malfunctioning? I assume so, but the rules should address it. Is the game clock paused while a program is restarted or computer rebooted, or does the side to move lose this time?

- The ICGA rules allow an operator to synchronize the computer clock with the game clock if they get out of sync. This seems to no be allowed. Is there any interest in allowing this? I am not in favor of rules which seem to not allow corrections for operator errors.

About the tie break, after you pointed out what seems like a serious flaw in the Sonneborn Berger, I am not in favor of using it. We could probably design something that is similar to the SB in that it factors in the final scores of each opponent, but in addition to giving positive positive points for wins, it should give negative points for losses. Surely there must be something like this already being used in chess tournaments?

Before the 2007 Culemborg tournament I asked Leo if there was a more detailed description of the rules than just the items in the tournament announcement. He sent me a document which I will post in the next reply to this topic. Maybe the people that frequent this forum can agree on some additions or changes to this document which can then be proposed to Leo.

-- Ed

[Ed Gilbert]

Rules for the open rapid draughts tournaments for computers in The Netherlands:

-the normal rules for draughts games dictated by the KNDB hold as far as they apply to computer draughts (operators need not be silent).
-the tournament is admissible to Dutch and foreign programs. It is an open tournament.
-operators bring their own computers.
-in each game both operators import the moves of the opponent program into their own program, and execute the moves played by their own program on an ordinary real 10x10 draughts board and press an ordinary chess clock afterwards.
-playing time per computer per game differs between tournaments, but for most tournaments it will be 20 minutes for 75 moves (all competing programs need to have an option for selecting a time schedule like this).
-official time played is indicated by the real chess clock, independent of the times indicated by both competing programs (typically these indicating times will differ enormously).
-after selection of the time schedule by the operator, the program and only the program decides the amount of time used for each move. Intervention by the operator is not allowed (for instance intervention via pressing a secret button). In case of an established offence the game will be declared a loss for the offending program. It may even happen that the offending program will be removed from the tournament (this scenario actually occurred once in an Olympic computer draughts tournament in London 1980).
-if after 75 moves both operators don't agree on the outcome of the game, the outcome will be established by the tournament referee. The referee is Jaap Bus, a first class draughts player.
-If possible each program plays once against each possible opponent; if the number of participants is too big, pairings will be established following the rules of the Swiss system.
-if at the end of the tournament more than one program made an equal top score, the top scoring programs will play one (or if necessary even more than one) playoff. Playing time in the playoff will typically be faster than the tournament time.
-the winner may call himself open champion computer draughts of The Netherlands and receives a true cup.



Glossary

KNDB : Koninklijke Nederlandse Dambond (Royal Dutch Draughts Federation)
Swiss system : a pairing system for tournaments with a large number of players and a small numbers of rounds.

[TAILLE]

Gerard,

I will probably not be at this tournament, but I am interested in your questions and improving the tournaments. I think if we want to see any changes in the rules we should decide exactly what we want and then propose them to Leo and the other programmers. In addition to your questions about tie break and playing more games, I have a few others:

- Under what conditions can the game clock be paused? Can a program be restarted if it appears to be malfunctioning? I assume so, but the rules should address it. Is the game clock paused while a program is restarted or computer rebooted, or does the side to move lose this time?

- The ICGA rules allow an operator to synchronize the computer clock with the game clock if they get out of sync. This seems to no be allowed. Is there any interest in allowing this? I am not in favor of rules which seem to not allow corrections for operator errors.

About the tie break, after you pointed out what seems like a serious flaw in the Sonneborn Berger, I am not in favor of using it. We could probably design something that is similar to the SB in that it factors in the final scores of each opponent, but in addition to giving positive positive points for wins, it should give negative points for losses. Surely there must be something like this already being used in chess tournaments?

Before the 2007 Culemborg tournament I asked Leo if there was a more detailed description of the rules than just the items in the tournament announcement. He sent me a document which I will post in the next reply to this topic. Maybe the people that frequent this forum can agree on some additions or changes to this document which can then be proposed to Leo.

-- Ed

Restarting a program : my understanding is that it is allowed but without any clock pause. No harm with this.

Synchronization of the clock : for me it should be allowed in order to limit the interaction of operator manipulation on the time allocated to the program, though this problem does not really exist for Damy.

SB : it is not quite satifactory but we need some time to try and build a new formula acceptable by all bodies. BTW the SB is not mentionned in the rule

I have another problem :
During the Damy-Cerberus game of 2007 Culemborg tournament it appeared that Leo Nagels made an operator error and, only several moves later, Leo Nagels detected that Cerberus position was not identical to the board position. Jaap Bus proposed to move backward till the beginnig of the problem to restart the game from that position. Of course I did not want to win a game on an operator error and I accepted to restart the game. The problems was the following :
1) We need some time to find where the problem occured in the previous moves. In the reality we made a pause on the Cerberus clock because I was imply in the search and I wanted to avoid penalizing Cerberus due to my own search.
2) By moving back in the game I was not able to tell Jaap Bus how many time was used by Damy for the last moves
3) I explained to Jaap Bus that Damy has lost some hashtable information and it was not that easy to restart the game in normal conditions.
4) Fortunetly I was able to change (in the middle of the game) the time available to Damy and we agreed to restart the game with a credit of two minutes for Damy.

What is your feeling about such operator error ?

[wellnesswrotter]

In Normal Game Situations a move is completed when the clock is pressed and you have to play that position UNLESS it is an illigal move AND the opponent wants it corrected. So only the last played move (and clockpress) can be re-taken.

In case of an Operator Error after a couple of moves, both parties should play the position as it is on the board. (So Leo had to put in the right position and let Cerberus calculate from there).

==jonah

[TAILLE]

In Normal Game Situations a move is completed when the clock is pressed and you have to play that position UNLESS it is an illigal move AND the opponent wants it corrected. So only the last played move (and clockpress) can be re-taken.

In case of an Operator Error after a couple of moves, both parties should play the position as it is on the board. (So Leo had to put in the right position and let Cerberus calculate from there).

==jonah

That sounds good sense but does it exist a programmer who accept to win a game on an operator error ? Of course you are right but as far as I am concerned I will accept any compromise in order to avoid winning a game that way. The number of programmers is very very low and our primary goal is to be able to challenge the other programs. I really prefer to lose a game rather to win on an operator error!

[wellnesswrotter]

But there has to be a rule about it. (and there is in the KNDB-rulebook).

For example: What if you're program can't go back in the game?
You don't want to win due to the operator error, but then you lose.

Ofcourse: if you can solve it friendly, do it.
(so take back the wrong moves, and set the clock back)

NOTE TO A PROGRAMMER:
Make sure you're program can
- Setup an position
- Setup clocktime
- Setup a game(notation)

[Ed Gilbert]

During the Damy-Cerberus game of 2007 Culemborg tournament it appeared that Leo Nagels made an operator error and, only several moves later, Leo Nagels detected that Cerberus position was not identical to the board position. Jaap Bus proposed to move backward till the beginnig of the problem to restart the game from that position. Of course I did not want to win a game on an operator error and I accepted to restart the game. The problems was the following :
1) We need some time to find where the problem occured in the previous moves. In the reality we made a pause on the Cerberus clock because I was imply in the search and I wanted to avoid penalizing Cerberus due to my own search.
2) By moving back in the game I was not able to tell Jaap Bus how many time was used by Damy for the last moves
3) I explained to Jaap Bus that Damy has lost some hashtable information and it was not that easy to restart the game in normal conditions.
4) Fortunetly I was able to change (in the middle of the game) the time available to Damy and we agreed to restart the game with a credit of two minutes for Damy.

What is your feeling about such operator error ?

This is a difficult situation. As you noted, it is not always possible to rewind the match back a few moves and then continue, because your program may not be designed to resume in this way. As you noted, you may not know how much time you had used for those moves that were undone. In my case, the only way I can change the clock is by resetting the count of moves played to 0. I could of course reset the clock to the remaining time left, and set the game moves to the number of remaining moves, but my search time algorithm varies during the game and takes into consideration that after a certain number of moves I am primarily getting results from an endgame database and thus do not need nearly as much time, so it would wrongly think it is at the start of a new game and not manage time correctly for the situation. While I agree with you that we should try not to penalize operator error, we also shouldn't allow an operator error to make the situation worse for his opponent. I think in these situations that if your opponent made errors causing his program's board position to differ from the table board, then you should have the option to request that he play from the position on the table, and only if you are willing to allow a rewind then you can offer this as an option. So in this case if you think you can rewind and resume the match without too much disruption to your game, you can offer this, but if not you can insist he play from the table board position. Does this sound reasonable?

-- Ed

[Ed Gilbert]

Restarting a program : my understanding is that it is allowed but without any clock pause. No harm with this.

Bert, didn't you have to reboot your pc during one of your games in Arleux? Were you allowed to pause the table clock during reboot and re-setup of the game state? Do you think pausing is the right thing to do?

-- Ed

[Ed Gilbert]

Synchronization of the clock : for me it should be allowed in order to limit the interaction of operator manipulation on the time allocated to the program, though this problem does not really exist for Damy.

I have thought about this some more and now I'm not sure it is a good idea. This privilege could easily be abused. I don't think any of currently active draughts programmers would abuse it, but looking at the chess world shows that there is a small percentage of people that will do whatever than can get away with to gain an advantage. Maybe it is better to not allow any clock adjustments during the game. I only wish there was a better method to keep the program clock in agreement with the table clock. The method you use in Damy guarantees that you will never lose on time, but you must waste a second or two on every move with the extra mouse click. The method I use of estimating the average operator time means that I could lose on time if I use more than I estimate.

-- Ed

[BertTuyt]

Ed, i also don't know the rules.
I thought, but it could be wrong, that i had to correct the situation with the clock running.
So in my case, and it happened two times, i didn't stop the clock.

Bert

[BertTuyt]

To be honest, i really don't like to do adjustments during the game.
I know (rumours) from the chess world, that some programmers also do on-line analysis of the game , and in this way add interesting positions to the hash tables.

Also i very much dislike, moving the mouse over the board during analysis, and "doing some imaginary move sequences".
Im sure no-one within the Draughts world has cheating modes built in their programs.

But to avoid any discussion, lets totally refrain from any major, non-standard, operator involvement.

Bert

[BertTuyt]

Ed,

and to give an answer on your question.
Think it is ok to pause the clock during a crash.
The other party then also has the advantage that the analysis continues, and more hashtables are filled in.

Bert

[TAILLE]

The method you use in Damy guarantees that you will never lose on time, but you must waste a second or two on every move with the extra mouse click.

That is not true. I do not really lose any second due to the extra click I have.

In fact I can very easily assured an almost exact correspondance between Damy clock and the official clock : let's suppose that I need t1 second in order to take into account the opponent move. That means the official clock is t1 smaller than Damy clock. Now consider Damy move. As soon as Damy chose its move, I need t2 second to play the move on the board and to press the button of the official clock and then, I need t3 more seconds in order to make the last click in order to inform Damy that the chosen move has now effectively been taken into account.
The more important times are t1 and t2 that have to be as small as possible in order to have the maximum time available when it is Damy to move. I introduced t3 only to assure an almost exact synchronization between Damy clock and the official clock. To reach an exact synchronisation between the two clocks I only need to assure that t1 is, on average, equal to t3 which is quite natural. In practice I typically start with an internal clock 10 seconds lower than the official one and, during the game, it is very easy to keep this value between 8 and 12 seconds.

You can note that it is almost impossible, with such procedure, to abuse it : let's suppose that you would like to force your program to take more time on the next few moves. In that case you would like to give more time to the program but, with the above procedure, you can gain time only by having t3 < t1. Because t1 has to be kept as small as possible you can see that the procedure do not offer such abuse.
If, on the contrary, you would like to see your program take less time for the next few moves you can of course force a high t3 but you can see that the time lost is surely definitly lost for all the game. Here an abuse is again impossible.

On the contrary a synchronisation directly on the clock may lead to very easy abuse. If typically you add 20 minutes to your internal clock you will force your program to take more time on the next move and nobody would like to see such abuse.

Damy approach is very accurate and, in addition, it is very efficient to avoid any abuse.

Do somebody see a problem with this approach ?