Page 13 of 16

Re: Use of tafl variants in real life tournaments

Posted: Sun Dec 29, 2013 7:01 am
by nath
The Vikings had much less opportunity to compare good strategies then we have today.
I like to draw back a line form the vikings, but I have to admit that want also promote Hnefatafl as a game itself and not only as a nice to know as part of the Vikings. And most assumptions that are made so far doesn't include any knowledge about balancing of the games.

To make a statement: I'd suggest to use Copenhagen Rules for official tournaments. The reason is quite simple. I'm to bad at this game to judge whether Copenhagen or Sea battle tafl 9x9 is better balanced, but we should agree on any basic set of rules to promote this game. We could take Sea battle tafl 9x9 now. That's true. But Copenhagen rules aren't much are difficult, if we add the rule that white has to break the permutal repetition. We even don't need the encirclement rule anymore (you can construct a situation, but it's nearly impossible to reach). Of course you have two extra rules, but I guess they aren't to difficult to learn (it's a feature like the rochade is in chess). The total complex of the rules is far more easy than it is in chess and on the combinatorial side Copenhagen Hnefatafl is the complex game. At the tactic part Sea battle tafl can catch up with Copenhagen rules. But the strategic part of Sea battle tafl 9x9 isn't important, you just play tactical.

I can also life with the choice to establish Sea battle tafl 9x9 as standard variant (in which our knowledge is weaker for far and that is combinatorial easier), but it would be a break regarding the the former promotion of the 11x11 board.
Even if I'd take Copenhagen Hnefatafl, the most important point for me is that we agree on any variant in common. I don't want Hnefatafl to become a game that forces you to talk about the rules before you can play a game.

Re: Copenhagen Hnefatafl

Posted: Sun Dec 29, 2013 9:43 am
by Hagbard
"Threatening player" cases are, typically, eternal checks. It is typically white who is the threatening party - the king struggles to get past black to corner or edge. In principle also black could be the threatening party, but I don't remember an example of this.

"Extra rule" cases are white draw forts.

Perhaps, in practice, the "extra rule" is sufficient to cover all cases of repetitions.

Re: Copenhagen Hnefatafl

Posted: Mon Dec 30, 2013 9:30 am
by nath
It is. It also covers the game between altti and me last tournament (that's the reason I didn't agree the draw), but it seems hard to test, because you have to retrace the whole game if you want to test whether the rule matters.

But I would suggest to change it to "white have to find another move". You can construct position in which it matters, but it's very unlikely to get such a position in a normal game and you get massive simplification of the rules. I don't want Hnefatafl to get rules you need a study a week before you understand them. And I needed your explanation to get this rule. We ugly need to change at least the wording, but I rather suggest to change the whole rule. It's very complex and not clear for new players.

Re: Copenhagen Hnefatafl

Posted: Mon Dec 30, 2013 10:07 am
by Hagbard
nath wrote:It also covers the game between altti and me last tournament (that's the reason I didn't agree the draw), but it seems hard to test, because you have to retrace the whole game if you want to test whether the rule matters.
No, it's known: the altti/nath game was caught by the "extra rule". The "first rule" would've taken an eternity to detect that specific case of repetitions.
nath wrote:But I would suggest to change it to "white have to find another move".
Worth a consideration.

Re: Copenhagen Hnefatafl

Posted: Thu Jul 24, 2014 2:11 am
by nath
I still can't see the huge advantage of black. I agree blacks position looks strong, but it's far beyond my ability to make it to a certain win. I guess the reason is also psychological. If we loose with black we usually see how white crushed us down and what defending moves we missed. We instantly we what kind of weakness we had. While playing white we don't get that instant feedback. Most times we played to passive or sometimes even to aggressive (which lead to a loss of important pieces). I guess quite often we got a weakness during the opening. Maybe the third or fifth move. I don't have a clue about the opening but I guess it's very hard to tell whether the move was actually a mistake or not. White winning variants are short and cruel. We can instantly tell. Black winning variants are much more complicated. The crazy thing is that we jump from one side to the other. Some guys are arguing for white, because we get simple straight wins to counter black mistakes (this also leaded to a higher amount of white wins between new players). We can obviously see that from the stats Hagbard created. Other side is to tell black is to strong, because if don't make a mistake he always wins. I can't see (against single player, but not in general) which color I should prefer.

Re: Copenhagen Hnefatafl

Posted: Tue Aug 05, 2014 10:23 pm
by Adam
Ultimately it's the long term stats that decide. Hopefully lots of players, old and new will join this years tournament, which will strengthen or weaken the case for Copenhagen accordingly. For the record, my money is still on Copenhagen.

Re: I dont know

Posted: Tue Aug 05, 2014 11:16 pm
by Adam
This would only be a problem if one was required to capture all pieces to win. That's why I was initially hesitant about the shield wall rule, as I felt it was superfluous. However, incorporated into Copenhagen with the edge fort, it has taken the game to new heights making end games far more complex.

Re: Copenhagen Hnefatafl

Posted: Wed Aug 20, 2014 5:17 pm
by Adam
In my last six copenhagen games with Crust white won all but one game, where black didn't spot the three sides+throne takes king rule.

So at least for me and Crust, black is harder just now. As Nath said, an individual players likelihood to win or lose as black tends to swing with improvements in skill and learning new strategies.

That's why I say we should trust Aage's statistics, but not personal experience. This years tournament will certainly give us lots of new data. I hope everyone who joins manages to complete all their games.

Copenhagen Hnefatafl

Posted: Fri Jan 09, 2015 10:12 am
by sqAree
Hello everyone!

This is sqAree from Berlin, first of all I want to say hi to the aagenielsen/World Tafl Federation community. Many thanks to Hagbard for running this great playing site and to nath for introducing me to the tafl game.

I personally like the Copenhagen variant a lot and can see its advantages over other tafl variants (of course it's also a matter of taste). Having said this, there have been more and more dicussions about the rules in the small chat lately and I'd like to contribute to the game by taking part in them.
I will refer to the rules as described here: Please note that I haven't read every single post in this thread, just point out if I write something that contradicts earlier posts. As for the translation I recently offered: I will work on it as soon as we settled on some final formulation of the rules.

Alright, let's get right into it:

I think rules 1,2,3,5,6 are well worded and need no further reconsideration.

About rule 4:

Contentwise fine, however it's a bit confused regarding redundance with rule 8. As there are multiple ways to win the game (for either player) we should possibly structure them so that no longer incoherent winning conditions for both parties are within one rule. Furthermore we might consider rearranging all the rules in way such that we have:
(A) Definitions and preparation (covered by rules 1,2,3).
(B) Moves and overall game mechanics (covered by rules 5,6,7 and also the first part of rule 8).
(C) Goals of the game respectively winning conditions (covered by rules 4,8,9,10). I know, usually (B) and (C) are transposed, but here the move mechanics are mandatory for the explanation of the winning conditions.
I'd suggest to start with the white winning conditions (corner escape and edge fort), proceed with the black winning conditions (king capture, drawfort and encirclement, although the latter possibly will become dispensable as a consequence of the eventual immobility rule and is only there to finish a game faster) and conclude with general outcome rules (immobility and perpetual repetition).

About rule 7:

At first I'd like to know a definition of "taflmen". Nath recently told me that the king cannot be captured by the shieldwall capture (which is consistent as otherwise a single king could die by being bracketed between 2 attackers on the edge), so there's something wrong here: Either "taflmen" are "attackers, defenders and king" or "attackers and defenders". However, the king may participate in the "shieldwall" but not as a "victim".
Of course we could rework the definitions now, but I simply suggest to add below: "The king may take part in the capture, either as part of the shieldwall or as the bracketing piece. The king cannot be captured using a shieldwall."

About rule 8:

Again contentwise fine, just the same as for rule 4, maybe split and rearrange. If immobility becomes a rule of its own, incirclement may stay as a special case (with the pure reason to shorten games as should be mentioned).
Also I wonder if the immobility rule has any significance in respect of a black loss. If black is indeed immobile, it seems that white can easily achieve a corner escape or an edge fort, because white has too few pieces to block every black piece in case the corners are secured.
What about this position (king:f11,black:e11,e10,f10,g10,g11,h11,white:d11,d10,e9,f9,g9,h10,i11). According to immobility white would win even though his king is surrounded. Of course white has a trivial win anyway by moving a piece, putting black into zugzwang and finally escaping with the king.
That makes me think of the following position (king:f11,black:d11,e11,f10,f9,g11,h11,e10,g10,g9). The king cannot move, fine. The black structure can never be captured, also fine. Now we distinguish between several cases (initially taking no account of loss by repetition of position):
Case (a): White doesn't have additional pieces. --> Undeniably black wins with either player to move, according to immobility.
Case (b): Black has no additional pieces, white has no more than 7 additional pieces. --> White cannot encircle the black pieces, so black will never be immobile. It's perhaps possible to put black into zugzwang but I guess white would lose pieces without uncaging his king then. That means, white would maybe win the game although his king certainly will never reach a corner or an edge fort.
Case (c): Black has a few additional pieces, white has no more than 7 additional pieces. --> As long as black's additional pieces don't get trapped, neither player can accomplish anything.
Case (d): Black has no additional pieces, white has enough pieces to encircle the black structure. --> This is an easy one: White blocks black and wins again by immobility despite his king being trapped.
Case (e): White has encircled the black structure, but black has additional pieces outside. --> 12 white pieces are not sufficient to savely block the black structure without losing pieces if black has enough additional pieces. But, say, black got one piece (beyond his 9) and white as well (beyond his 8). Black continually tries to threaten the black heap but the one defender is always preventing a capture.
A funny position would be (king:f11,black:d11,e11,g11,h11,j11,e10,f10,g10,h10,i10,f9,g9,white:c11,d10,e9,f8,g8,h9,i9,j10,j9). Black would be immobile if he hadn't constructed a black fort.
I will take up on this later.

About rule 9:

As earlier mentioned in this thread, consider a position where white's only piece is the king and black has secured all corners using 3 pieces each. Technically white would win, if the king has contact with the edge, because he is able to move and cannot be captured really. Well, as for this rule I do understand how it is supposed to work, it's just the wording that is ambiguous in my opinion. We would need to clarify if it's "impossible for the attackers to capture him in the next move" or "impossible for the attackers to capture him in the course of the game" or "impossible for the attackers to capture him after any amount of moves assuming that white passes all the time". Of course only the latter would be right and it's also my suggestion for a reformulation, because when reading the actual rules as is, every of the above interpretations could come to one's mind.
However, there is still an issue, namely the cases where the king is part of the fort structure, because this situation is supposed to be a black win. My final suggestion for a reformulation is as follows: "The defenders win if the king has contact with the board edge, is able to move, it is impossible for the attackers to capture him after any amount of moves assuming that white passes all the time, and it is impossible for the attackers to render white immobile assuming that the king is forced to move at least once."
I hope this is correct. Anyway, it's anything but descriptive, so an alternative would be: "The defenders win if the king has contact with the board edge, is able to move, and it is impossible for the attackers to win the game if they control their pieces plus the king while the defenders don't move."
At least this would save us defining precisely what an edge fort essentially is.
In any case I recommend adding a few more examples of edge forts. Really, not a single formulation of rule 9 is intuitively accessible so far.

About rule 10:

To start with a few questions:
(Q1) Does this rule apply for both players?
(Q2) If yes, who is the "threatening player", how is he determined?
(Q3) Is this a sudden death or is the move that would lead to the repetition just banned? (of course the answer doesn't matter as in case of no sudden death the immobile rule would do the job)
My further reasoning will depends on the answers. I think, this is the point where it comes down to defining what copenhagen Hnefatafl is all about.
Everyone would agree that white's goal is to let his king escape to a corner. Now what should black's goal be? It can be "capture the king" or the contradiction to white's goal.
Given that it's the former, the idea of forts pops up immediately: A white structure that can live autonomously prevents the king capture. If I recall correctly those forts are counted as draws in Fetlar Hnefatafl which is quite reasoned; neither player can reach his goal. Now we distinguish between edge forts and usual forts, probably because reaching the edge is somewhat more of an escape then a draw fort in the centre of the board. Alright, so edge forts win for white and other forts should be truly worse. It can't be a draw though because white would have a trivial draw from the starting position then. Concluding the non-edge forts count as losses for white.
So far, so good, this was just a small recapitulation of what I think the chain of thoughts was when designing copenhagen tafl.
Now let's pose this question once again: What should black's goal be?
Let's say it's "capturing the king". --> Corner escape (white win), king capture (black win). Edge fort (white win), but draw fort (black win)? Both of the forts accomplish two things, that is, preventing the king from being captured and getting rendered immobile. In this case both forts should count as a draw actually, because no player has a chance to reach his goal. Even if we put edge forts aside; there are still cases where black simply can't win by capturing the king, but rather by encirclement. Not to mention the positions I described under rule 8.
What if black wins by ensuring white doesn't reach his goal? Would be my first choice when designing a game, because it just sounds too logical. Note that there is no contradiction to any of our working copenhagen rules; every case that is covered by them would have the same outcome under this assumption.
Now we have cases (b), (c), (d), (e) (see my comment for rule 8) as well as the case where black has secured each corner and both players have very few pieces left (as described in the small chat, let's call it (f) for references) plus maybe the black fort (g). Let's now start applying the gathered points and rule 10 to those cases:
(b) No matter if white got at least 4 attackers or not, the outcome will be decided by which player goes out of non-repetetive moves first (assuming Q1 is answered yes). Since black has no additional pieces and needs to sustain his structure, he will most likely lose.
(c) Actually the same as (b), the outcome will depend on Q1 and the number/arrangement of each player's pieces.
(d) Nothing would change, it's still a white win. I find this kind of counterintuitive though.
(e) Again the outcome depends on Q1. If only white is not allowed to repeat positions, it's a black win. Otherwise Q2 gets interesting.
(f) Technically white would lose according to the draw fort rule if he has at least 4 defenders and can't build an edge fort. But it might take so many moves that it's impossible to finish the game in practice. If he has 3 or less defenders, maybe the perpetual repetition would be effective eventually (Q1,Q2).
(g) Black can only move one piece back and forth. Once again Q1 and Q2 will decide.
So, quite a lot depends on Q1. If the answer is "yes" the example positions are a mess to deal with, because the result will almost always be revealed only after a long time. And even if we decide to allow draws (it's still not in the rules and admittedly I'm against draws) there remains a lot of work to decide the outcome arbitrarily for those border cases. That's why I propose to make it possible only for white to lose to the perpetual repetition rule. Of course this can be abused in the midgame by repeatedly threaten the king and winning as black without capturing him. But I think this might actually be a nice tactic.
So, let's look how the cases behave under this ruling:
(b) Black wins if he manages to uphold his structure. Success might be difficult to prove, but not impossible. I believe that white wins comparatively fast if he has a win, so an "after x moves black wins"-rule could help to detect the outcome. In chess it's standard that the umpire sets a match's result if one player does not try to win (in his opinion). Either of those rules could be applied here.
(c) Black moves his other pieces and wins because white will have to repeat a position at some point. Seems fair to me. Get a referee to end the game faster.
(d) Although this seems counterintuitive it's not really arguable that it's a white win. This might be the only situation where white does not win reaching one of his main goals. Though as if this position will ever occur anyway..
(e) Would be a black win analogous to (c).
(f) Nath mentioned this somewhere in the small chat. If white has 3 or less defenders, he loses (just a shortcut for the perpetual repetition rule). If he has 4 or more, however, I can't see how black can secure every corner and let at a max 2 squares per edge for white while he has too few pieces to win the game.. Let's just say, analogous to (c) and (e) white loses after a fixed amount of moves (he should get a chance to build that edge fort at least) or if an umpire decides so. I am deeply convinced that it should be a black win, because white hasn't accomplished anything: No corner escape, no edge fort, no encirclement of black pieces.
(g) Interestingly this would be a black win, obviously. The black fort's charme is quite decreased because now he wins in the other cases as well. What a pity..
All in all, it seems like a rule analogue to the Fifty-Move-Rule in Chess (maybe a different number, depends on how many moves it will usually take to build an edge fort, and a black win instead of a draw) would resolve all cases. I agree that chess is a bit flawed because there are positions that would be a win without the Fifty-Move-Rule, check ... forced_win, a position with white to play and force mate in 549 moves. Then again we have endgames like Bishop and Knight against King that win within 50 moves but it's okay to punish a player if he fails even once. We might want to compute a number x, such that an edge fort - if possible - can always be built in x or less moves.

Alright, I was writing for hours and am extremely tired, but I said anything I wanted to say - for now. I guess my text will be flawed as hell, please bear with me, I'm only a beginner.


Re: Copenhagen Hnefatafl

Posted: Thu Jan 22, 2015 10:50 am
by Adam
Great to see new enthusiasm and such thoughtful analysis of the rules SqAree. I need to find more time to read and respond fully, but one point that jumps out for me is your point about what black and whites respective goals are. I agree that this is fundamental to refining the rules and resolving our grey areas in the game. As it stands blacks goal is clear, to capture the king. Anything less is failure as the siege simply continues. We could indeed change this to 'prevent the king from escaping''. This would make blacks job simpler. I would like to clarify one thing, we should refer to edge forts as 'exit forts'. That way it is clear that it is not a stalemate position. The boards impenetrable edge is breached by construction of the exit fort. I like to imagine that the fort buys them time to dig an escape tunnel!

Black covering corners is absolutely a draw position and not a win, given the present agreed goal for black. The effect of black only needing to prevent escape would make this draw position a black win, providing white is down to 3 soldiers making an exit fort impossible, or that the edge is not reachable. I'm not certain that this will improve the game, only simplify it and favour black. Maintaining a siege doesn't conclude the situation. Perhaps we need stronger language, like: blacks goal is regicide!

Discuss :)