Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> General Discussion >> Repetition rule - clarification
        
(Message started by: lazarov on Nov 22nd, 2014, 7:22pm)
Title: Repetition rule - clarification
Post by lazarov on Nov 22nd, 2014, 7:22pm
Hello,

I have a question about repetition rule:

AFAIU it applies when after "player A" produces the very same board position, i.e. placement of all pieces, 3 times:

1) Does it have to be 3 consecutive times or any 3rd time, perhaps with lots of totally different positions in between?
2) How is it enforced by the software on arimaa.com from human perspective? Would I lose a game if it happens to me, or I would not be able to make that move in a same way as I can't move a frozen piece?
Title: Re: Repetition rule - clarification
Post by browni3141 on Nov 22nd, 2014, 11:30pm

on 11/22/14 at 19:22:27, lazarov wrote:
Hello,

I have a question about repetition rule:

AFAIU it applies when after "player A" produces the very same board position, i.e. placement of all pieces, 3 times:

1) Does it have to be 3 consecutive times or any 3rd time, perhaps with lots of totally different positions in between?


Any third time.


Quote:
2) How is it enforced by the software on arimaa.com from human perspective? Would I lose a game if it happens to me, or I would not be able to make that move in a same way as I can't move a frozen piece?

The client will tell you that your move results in threefold repetition, and won't let you play it. You won't lose unless you don't have enough time on your clock to undo and make another move.
Title: Re: Repetition rule - clarification
Post by lazarov on Nov 23rd, 2014, 1:50pm
Thanks!
Title: Re: Repetition rule - clarification
Post by minionofmidas on Nov 27th, 2014, 1:28pm
And how is this rule enforced when playing an informal live game? Ignored, I suppose - I doubt these situations occur all that frequently?

(Brand new here and brand new to the game, btw. Hi.)
Title: Re: Repetition rule - clarification
Post by browni3141 on Nov 27th, 2014, 3:20pm

on 11/27/14 at 13:28:11, minionofmidas wrote:
And how is this rule enforced when playing an informal live game? Ignored, I suppose - I doubt these situations occur all that frequently?

(Brand new here and brand new to the game, btw. Hi.)

By "live" do you mean offline? If so then I don't know if there are procedures in place like there are for chess's threefold repetition. I assume the opponent would be required to make a claim if you made an illegal move by repeating the position for a third time. It's not too hard to keep track of repetition in nearly all practical cases.

Edit: The repetition rule isn't applied rarely, but almost all the time it is when the opponent directly undoes your move, which is really easy to keep track of. Cycles of two moves are rare, but I've had them a couple of times. Cycles longer than two moves are extremely rare.
Title: Re: Repetition rule - clarification
Post by minionofmidas on Nov 27th, 2014, 3:38pm
Yeah, I promptly came upon such a directly-undoing situation in the next game after I wrote that, and understood how that rule will force you to make a suboptimal move. :D And yeah, a situation like that one would be easy to spot in a non-computer game.
Title: Re: Repetition rule - clarification
Post by Fritzlein on Nov 28th, 2014, 7:12pm

on 11/27/14 at 15:20:51, browni3141 wrote:
Edit: The repetition rule isn't applied rarely, but almost all the time it is when the opponent directly undoes your move, which is really easy to keep track of. Cycles of two moves are rare, but I've had them a couple of times. Cycles longer than two moves are extremely rare.

Actually, if I recall correctly, I counted once and about half of all games don't even have an opportunity for either player to undo a move the other player makes.  So I believe it doesn't come up "almost all the time".

Furthermore, even if undos were allowed, I estimate only 1% of games would be drawn for that reason.  Usually there is a way for the aggressor to make progress in spite of having moves undone.  Thus it wouldn't be catastrophic to have no rule whatsoever about repetition, allowing a small number of draws.

Finally, for face-to-face play, the "no undo moves" rule will, in my estimation, suffice to make 99.99% of games have the same outcome as under the full server rules.  Since it is easy to enforce and adequately accurate, there is essentially no problem to keep repetition out of ftf games, thus avoiding draws.
Title: Re: Repetition rule - clarification
Post by browni3141 on Nov 28th, 2014, 8:52pm

on 11/28/14 at 19:12:37, Fritzlein wrote:
Actually, if I recall correctly, I counted once and about half of all games don't even have an opportunity for either player to undo a move the other player makes.  So I believe it doesn't come up "almost all the time".

Perhaps I was unclear, because you seem to have misunderstood. I was trying to say that in nearly every case the repetition rule is applied, it is in the case where the opponent may directly undo your move, as opposed to cases where two or more moves are played before a repetition occurs.

Quote:
Furthermore, even if undos were allowed, I estimate only 1% of games would be drawn for that reason.  Usually there is a way for the aggressor to make progress in spite of having moves undone.  Thus it wouldn't be catastrophic to have no rule whatsoever about repetition, allowing a small number of draws.

I'm not sure how you came up with 1%, but I agree that it would be a small number.

Quote:
Finally, for face-to-face play, the "no undo moves" rule will, in my estimation, suffice to make 99.99% of games have the same outcome as under the full server rules.  Since it is easy to enforce and adequately accurate, there is essentially no problem to keep repetition out of ftf games, thus avoiding draws.

I'm not sure if this is still replying to me, but I agree. I would require my opponent to keep track of repetitions regardless of how long between them, but I can't complain if someone else doesn't want to do that :) In a tournament situation I would require all repetitions to be tracked also, but for casual games who cares as long as you're enjoying playing Arimaa?
Face to face prohibiting only direct reversals as opposed to any repetition of position won't change optimal play very often.
Title: Re: Repetition rule - clarification
Post by half_integer on Nov 29th, 2014, 3:32pm
Maybe if I thought about it longer it would be obvious, but is there any difference between a "no second repetition" rule and a "no third repetition" rule aside from the case below?

The only difference I see is that it allows a player who changes strategy and needs to progress back through a prior state to pursue that strategy to do so.  This would usually be reversing a previous move to reposition pieces but could rarely require passing through a much older position in pursuit of a strategy.

In the usual direct-undo by the opponent, it simply allows the opponent to indicate their desire to undo the move, allows you to decide to pursue a different direction or not, and if not forces them to choose a different response.

I'm not questioning the rule; I think it is well written to cover longer sequences.  I just want to know of any special cases when explaining it to others using the simpler form above.
Title: Re: Repetition rule - clarification
Post by Fritzlein on Nov 30th, 2014, 10:19am

on 11/28/14 at 20:52:42, browni3141 wrote:
Perhaps I was unclear, because you seem to have misunderstood. I was trying to say that in nearly every case the repetition rule is applied, it is in the case where the opponent may directly undo your move, as opposed to cases where two or more moves are played before a repetition occurs.

Ah, yes, I did misunderstand.  Now I see what you meant.


Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1!
YaBB © 2000-2003. All Rights Reserved.