Repetition of moves: some dubious positions

Tafl rules
User avatar
Hagbard
Posts: 691
Joined: Sat Mar 14, 2015 6:07 pm
Location: Copenhagen
Contact:

Re: Repetition of moves: some dubious positions

Post by Hagbard » Sat Sep 07, 2019 11:56 pm

Hagbard wrote:
Sat Sep 07, 2019 10:59 am
First situation:
You succeeded in finding a case where the repetition is noone's fault, both sides are only doing forced moves, noone is driving this repetition.
I think I was too quick with this one.

It does not matter, what's on the other side, when a piece attempts to bypass its opponent's blocking. It could be a certain win in 2-3 moves like here, escaping to freedom also like here, or whatever else.

So the king attempts to break out, causes repeating and loses.

Ytreza
Posts: 20
Joined: Tue Aug 27, 2019 5:56 pm

Re: Repetition of moves: some dubious positions

Post by Ytreza » Sun Sep 08, 2019 11:27 am

I appreciate that your examples support your suggestion that all repetitions could be punished to white.
My goal here is not so much to defend this suggestion but to show that the current statement of the repetition rule is dubious. In the first situation, it looks not fair to punish white (and neither black). Of course, stating that white always loses on repetitions is a way to solve all the issues I showcase here.
It would not be fair to leave white defenseless
Well it depends how you formulate the win conditions. If you state that the goal for black is to prevent white from putting the king in safety, repeting the moves is a way to do it. After all, white is besieged, and a besieged opponent loses if he doesn't win quickly enough. This was my suggestion about a "starving rule". Actually one could even extend it and include repetitions in "if no capture is made in 100 moves white loses". This indeed implies that white loses on repetitions.
I think I was too quick with this one.
Isn't this a clue that the rule is unclear? Even an experimented player like you struggles to interpret it correctly :)
So the king attempts to break out, causes repeating and loses.
So back to this formulation in terms of who is blocking. I see 3 issues with this one.
1/ This formulation is not equivalent to the current formulation of the rule. See the position below:
Diapositive6.PNG
Diapositive6.PNG (41.97 KiB) Viewed 19302 times
Who "causes" the repetition? Well, white is not threatening anything, he just tries to not die. On the other hand, black has clearly the choice. So my intuition tells me that black causes the repetition. But it is black who is blocking white.

2/ This formulation is not fair to white.
As already mentioned, the goal of the game for white is to find open paths and the goal for black is to block white. So this formulation leaves white defenseless in many cases. Like in the previous position. It looks paradoxical to reject by virtue of fairness the simplest rule saying that white always loses, while the alternative statement is unfair to white.

3/ Up to now, I actually have no problem with the formulation in terms of who is blocking. I don't care if it is fair or not, the game is already asymmetrical to start with... The only issue is that it is actually not the statement currently written in the Copenhagen rules, but ok. However, it has a major issue when the repetition involves more than 2 stones (repetition cycle, or board position repetition). In this case, both players could block something alternatively. Look at the following position:
Diapositive1.PNG
Diapositive1.PNG (42.22 KiB) Viewed 19302 times
White has almost achieved an edge fort, while black is in a quite good position. The repetition is done in 4 steps shown below:
Diapositive1 - Copie.PNG
Diapositive1 - Copie.PNG (151.2 KiB) Viewed 19302 times
Now imagine you're a referee at a tournament. Black stops the timer, calls you, and says:
"Mister referee, white is violating the repetition rule! He wants to get his king in the edge fort, so I block with move 1. Then he plays move 2 and threatens to capture my pawn. Since I don't really care about the capture, I play move 3 to seal the corner. Then white plays move 4, threatening to enter again inside the edge fort. So I block with move 5. Then white threatens to capture again with move 6, but again, I don't care, so I play the innocent move 7, and white threatens to enter again! I'm the one blocking the king with move 1 and move 5! It's the king who seeks the open path!"

But white replies:
"Mister referee, with all respects, I don't agree with my dear opponent. It is me who keeps blocking black. Indeed move 1 threatens to capture my pawn in c10, so I block with move 2. Then move 3 threatens to capture my pawn in c11, so I block with move 4. Then move 5 threatens to capture my c11 pawn again, so I block with move 6. And finally move 7 threatens to capture my pawn in c10, and I block with the king."

Now you, referee, must take a decision: who should break the repetition cycle?

User avatar
Hagbard
Posts: 691
Joined: Sat Mar 14, 2015 6:07 pm
Location: Copenhagen
Contact:

Re: Repetition of moves: some dubious positions

Post by Hagbard » Sun Sep 08, 2019 12:40 pm

1) If ignoring the white piece on c11 which would capture c6 so that the king escapes and wins,
if c11 were not there, then white is entrapped and repeats c5-d5, blocked by black, and white loses.

3)
Complex. The king attempts to enter the fort and is blocked. Black attempts to destroy the fort and is blocked.
So when the pattern is considered undefined, a draw.

nath
Posts: 65
Joined: Mon Oct 09, 2017 9:34 pm

Re: Repetition of moves: some dubious positions

Post by nath » Sat Sep 21, 2019 7:10 pm

You are making very suspicious judgements. We could try to clutter the rules with a lot of extensions, which are just rulings of single cases. Imho that is just awful and will never lead to understandable, easy or concise rules...and I'm not sure if they will be formal correct that way like ever.

I made the simplification request along with several members of the Berlin community - most notably sqAree who made a serious attempt which lead to minor rule change - several times since 2013 already.

Let me quote my last attempt, that was left unanswered:
nath wrote:
Wed Nov 21, 2018 2:21 pm
Is there a real reason we want those draws anyways? 8/8b seem to strongly contradict rule 10. Why not simply declare this a black win? We did that in in various offline games, since that removes the dire need for an umpire all the time to decide with rule takes precedence (which is highly subjective anyways).
I don't understand why we need to discuss every of those examples in detail, when it's since 2013 so obviously clear that the rules are broken. I just plead for a systematic cleanup of this and just for a an addition that is just taping over jet another crack.

Regards
nath

Ytreza
Posts: 20
Joined: Tue Aug 27, 2019 5:56 pm

Re: Repetition of moves: some dubious positions

Post by Ytreza » Sun Sep 22, 2019 2:24 am

I agree with nath. That's the point of all my examples: showing tricky exceptions. I also agree on the solution: white should win or lose. The balance is not expected to change much in any case as repetitions are very often "white's fault". It's simple, elegant, easy to understand, easy to implement numerically and somewhat consistent with the "viking spirit": win or die.

I'm also more and more convinced that black should win whenever white cannot win anymore, i.e. all corners are closed and there is no space on edges to build an edge fort or not enough remaining defenders. Playing endgames as black is often very boring (see e.g. my WTF tournament game against jrton80) and white's play only rely on the hope that black will make a mistake at some point... In many games it is clear that the winning chances are 100% for black at move 60, but still the game is 150 moves long... Also, the rule saying that there is a draw if no capture is made in 100 moves can be at black's disavantage: it takes a very long time for black to close the barrier and reaching 100 moves (i.e. only 50 displacements of attackers) seems very possible. I don't like that white has the possibility to negociate a draw by just retarding (annoying) black while in a completely lost game. Sounds like bad sportsmanship imho.

Both issues of repetitions and long games can be solved in one line: white wins or loses.

A minima and given the discussion we had in this thread, I think rule 8 should be reformulated as "the player who is not blocking must find an alternative move", which I demonstrated is not equivalent to "the player who causes" or "the player who has the choice". Still, speaking of "blocking player" only solves repetitions involving not more than 2 moves, and in a rather complicated way.

Also, I see that rule 10 has been reformulated. The statement "it is not possible to end the game" is again quite dubious imho. How can one be sure that there is no variation ending the game in 500 moves? Behind this there is actually simply the draw by agreement and I think the formulation should refer to that explicitly. Then what about previous rule 10?

unhandyandy
Posts: 39
Joined: Fri Dec 28, 2018 6:39 pm

Re: Repetition of moves: some dubious positions

Post by unhandyandy » Mon Sep 14, 2020 9:25 pm

I agree with Ytreza
I'm also more and more convinced that black should win whenever white cannot win anymore
and nath, that the rules should be clear.

I think White should be forbidden from repeating a position, but not Black.

This would also be in line with the theme: White is trying to escape, time is against him, if he gets into an infinite repetition he loses.

User avatar
Hagbard
Posts: 691
Joined: Sat Mar 14, 2015 6:07 pm
Location: Copenhagen
Contact:

Re: Repetition of moves: some dubious positions

Post by Hagbard » Thu Apr 22, 2021 10:52 pm

October 2020 Adam Bartley started a discussion with Mario Aluizo ("casshern") and myself about repetitions in Tafl, and we came to the conclusion that the repetitions rule can be simplified, so that repetitions are always a loss for white (the defenders).

I searched the full archive ten years for games ending in black repetitions and found none.

Repetitions being always a loss for white has also earlier been suggested by Ytreza, Nath, OdinHimself and Unhandyandy.

So the repetitions rule will be:
Rule 8: Any perpetual repetition results in a loss for white.

Post Reply