Adjudicate games with program? Constructive Input welcome

[rich beckwith]

The idea of adding a new rule to permit using a computer program to adjudicate overrun games was discussed at IL, but could be applied to other tournaments as well, including a National. Laptops are commonly found at tournaments now. I presume the procedure (if adopted) will work as follows:

For a position that cannot be obviously adjudicated by sight, the referee will enter the position into a program, such as Kingsrow or equivalent, preferably one with an endgame database. I would also suggest a second person (co-referee or other playing rules committee member) verify correct entry and move of the entered position. If position is not already in a database, then a minimum think time (2 minutes?) is recommended to help ensure the position doesn’t fool the program into a wrong verdict.

Secondly, criteria need to be set to indicate the cutoff between a draw and win. While most positions are “easy” for a program, there may be a critical 6 x 6 position where the computer gives an indecisive result, even after allowing additional think time. Perhaps a rule should state a minimal evaluation after 5 minutes of +60 (or whatever the optimal number is) is required to declare a win.

My analysis:

No matter what system you use to deal with overrun games, there will be flaws. The question therefore is, “What is the optimal solution?” One critique of adjudication: what if someone knows they have a published win but doesn't know how to win, and refuses to move so the computer (or other humans) will adjudicate? Hopefully players wouldn’t resort to such tactics, and stalling would be harder to do at a tournament where clocks are used. In contrast, many people don’t like "move-a-minute" since someone can easily lose a game through skittles for which they earned a draw (or win).

I think a program is accurate more quickly than humans on most positions, esp. when there are few pieces on the board so that the program knows the result by its database. Also, sometimes your best human adjudicators may be tied up playing the run-over games.

Another option is to do move-a-minute for 10-15 minutes and then adjudicate if game by program if the game is still incomplete, but there is still some degree of skittles here.

I would also welcome input from anyone who knows what other federations/games do on this modern topic.

[Mr. Checkers]

Rich lets assume that all of this worked out as planned. Then who would be given credit for the results--the program or the checker player? The "human factor" has made checkers what it has been for many many years. To turn that over to a program would certainly not be acceptable to those that view our game in that way. "Inky"

[Alex_Moiseyev]

I would also welcome input from anyone who knows what other federations/games do on this modern topic.

It's easy and simple - use clocks and set time control 30 (or whatever) minutes per game to each player. Flag failed - game over. In this case round will be finish in 2 hours.

You also can use "Fisher clocks" with increment time.

Alex

[Patrick Parker]

i think clocks should be used even though i dont really like them
i wouldnt want to isolate anyone .....but lets face facts
if i can watch 11 year old kids at a chess tournament use clocks i believe everyone thats an adult in the checker world may also be as capable.....
maybe we can try .....clocks in all master division tournaments .....not as a rule but something to try to get started .....encourage it to be done....
this isnt always needed but i hate when a player takes 45 minutes on a move in a losing position with 16 pieces on the board............im not joking

[Ingo_Zachos]

In chess, we always use clocks.
The time rate is usually at 40 moves in two hours and 30 minutes for the rest of the game.
There r special rules for ending the game at the second time intervall (the 30 minutes above), called
"quick play finish", ie:
A player can always stop the clock if he has less then 3 minutes to ask the referee for a draw.
The referee then has to declare the draw if the position is a winning position for the player that asked for a draw or if it is a clear ("theoretically") draw.
If the referee declines the draw, the game will be played until the end.
In no case the referee decides "loss of a game"

The use of chess programs and endgame databases has signifiacantly reduced the number of questionable decisions by the referee, and I see no reason not to use them in checkers.
I would recommend someting like 30 minutes or an hour for a whole game in weekend ty's and maybe 20 or 30 moves in one hour with 30 minutes for the rest in a bigger event like the WCDF Qualifiers or the US nationals that span over a week.

Greetinx from sunny and bright Dortmund, Germany

Ingo Zachos

P.S:

In chess so called "Blitz" tournaments r also very popular:
Each player gets 5 minutes for the whole game.
Most counties in Europe hold a special national Blitz ty, and also team events at Blitz.
It is a lot of fun, and very recently we even had a World Championship at Blitz.

[vhabgood]

some other things to consider with using programs... 1: most players can not play the position as well as the program. if a tough ending such as 'bowens twins' for instance, if the opponent does not how to win it and can not prove that he/she is making progress, than the game should be considered draw. 2: there are errors in some databases! albert tucker told me about how he got into a 4vs 4 ending (with a program that has 8 piece complete database) and he got into a position where the computer and him just moved back and forth over and over but the evaluation of the position was a win! 3: if the ending is within the database, the computer will respond right away from memory, thus eliminating the need for a timer as previously suggested. Finally, I would just like to add that this a gentleman's game, and both players should be respectful of one another.
---Victor C. Habgood II

[Alex_Moiseyev]

Richard, I find solution !!!

Also simple and absolutely realistic, which resolve all problems and should satisfy everyone, not complicated procedure which can be use in any state.

Referee or tournament directors should always have one tournament clock, just one only ! I don't think this request is too crazy. If two players don't get an agreement and round time is over, referee should set a clock giving each player 3-5 minutes, and then they should play the game to the rest. (In Russian checkers blitz was - 3 minutes to each player, in 10x10 - 5 minutes to each player).

In this case no programs or any other kind of judgement involve - our main goal and target.

Think about ...

Sincerely,

Alex

[tommyc]

The ultimate solution is everyone use a clock,i would hate to think what any chess player reading these posts might think of us ,the word amateur comes to mind..

In the final analysis we must put things to right sooner or later,altho there does seem to be a resistance in some quarters in the use of clocks. Can we justify the "none" use of clocks"No" i dont think we can in the 21st century.

[Pedro Saavedra]

One other chess option (and one available in most digital clocks) is the five seconds lag rule. The first five seconds when it's your move do not count. That means that if a game is an uncomplicated draw a player would not play on hoping for the opponent's flag to fall as the other player should draw without using time.

By the way, the rule of being able to ask the director to declare a draw involves (at least in tournaments I have played) no losing chances. If the position is known to be a draw, but there are many ways in which a player can go astray the game would be played on in chess.

[tommyc]

You have missed the point Pedro,we are talking here running a Ty without clocks/
Do you have a solution?

[Gene Lindsay]

I think it is more fair to have the game ajudicated when time is up than to be forced to play blitz checkers and lose a drawn or won position because you didn't have time to look the position over. At the 1994 National Tinsley
made the comment he would never play in a tournament with the move a minute rule, he would agree however to computer ajudication.

Gene Lindsay

[Alan Millhone]

I fully agree with Gene that 'move a minute' is not Checkers. True, both players have to play under the same format in 'move a minute' but after both players have played a hard fought game to then settle the reuslt by moving quickly does not 'float' in my book.

Computers and Checker programs are here to stay. More and more players (including a 58 year old one like me) now bring along a notebook computer to tournaments.

Currently there is a rules committee set up by Mr. Hugh Devlin to study the rules and attempt to standardize the game where ever one plays. Personally I would love to see clocks used at all tournaments to also standardize and elevate the stature of the game.

Sincerely:
Alan Millhone, President
American Checker Federation

[Alex_Moiseyev]

'move a minute' is not Checkers.

Probably, but still better than judication by human or program. The game between two players cannot be terminated - this is the only rule which we have to follow.

If you don't have a time to play, even move-per-minute - use a coin. In my opinion It is still better than use a program or other player judication.

Alex

[Mr. Checkers]

Fellows what ever happened to the idea that if the game is not won by the time the clock runs out that the game automatically becomes a draw--regardless of the fact that one player may or may not have some kind of theoritical advantage for whatever reason? I see far to many flaws in saying that the computer says you were in a lossing position therefore you lost--then a few years down the road another computer finds an error in that first computers analysis; therefore you were the winner after all. That was really tough luck Charlie and we are sorry for our mistake. And looking at it from a different perspective if a player has a winning position and can not bring that about then lets give credit to the underdog who was foxy enough to prevent him from getting a win--then send the other guy home to study more in depth how he could have won his game. 2 checkers against 1 checker is not always an automatic win for the person having the 2 checkers. And whos games are you going to adjudicate? Just the best players? And who is responsible enough even to be adjudicating them in the first place--the good players who went out on the town the night before? And when will this be done? Most players don't hang around after a tournament is over. And who's computer program is error free? As we all know computers are only as good as the knowledge programed into them. I don't see this adjudication process (with computers) as being the solution this the situation, but I do see it as a mechanism to complicate the situation even more. There are far to many human variables involvled here (to many ifs) for this to be the panacea--so take the draw--go home and study. Someone once said many years ago that a good checker player never makes the same errors twice. "Food for Thought". "Inky"

[Ingo_Zachos]

:lol: Fellows what ever happened to the idea that if the game is not won by the time the clock runs out that the game automatically becomes a draw--regardless of the fact that one player may or may not have some kind of theoritical advantage for whatever reason? ...

That's the idea of adjudicating a game.
So you finally got the point!

The question is: how can one determine if an advantage is only "theoretical" or "decisive".
Both players may disagree about that.
What then ?
And here a computer or a strong player may help to decide the matter more profound than any cat can do (plz forgive me, but I am not better then a grandmaster or kingsrow. But so are you!).
BTW: As I understand it can only help to call it a draw, not a loss.
If it says, it is a loss, the player may still try to prove a draw in his remaining time. That is only fair.

There is also another grave foul play in your argument:
Once the time has fallen it IS a loss.
One has to ask for a draw before he forfeits the game.
That is only fair, as the other player can not be blamed if I decline to play, but rather unfair refuse to make any move.

BTW:
Which other rule do you propose ?
To allow a player never to move at all if he desires so ?
Why do you think the 5 minute count was invented ?


Happy Easter Holidays from Old Europe,

Ingo Zachos