Anyone want to join cryptoguru's Speedy Scotch Tablut tournament to help him celebrate becoming a Brain Rook? Join NOW!

cryptoguru: Personally I would like that. But then you and I are mathematicians. Most people I suspect will be happy to take it on trust though, as you suggest, it's easier to take something on trust when it's in the open (even if you don't understand it).

The previous system seemed to me to be done by a series of lookup tables which was sensitive to the difference between to the players and applied fixed increments/decrements in 200 BKR bands. What is done now I haven't had enough time to notice.

mikkyT: It's a good job that I know you personally and that (a) I know when you're joking and just trying to be provocative, and (b) I can come round to your office desk with a large baseball bat and give you a moderating you won't forget in a hurry. ;-)

Work within the system guys. And be nice. :-)

redsales: Tablut theory definitely exists - which is why I'm writing a book about it. And Tablut theory is developing all the time. Or maybe the theory is "recovering" instead, and we are "rediscovering" what ancient north European players knew for centuries. The double fortress draw that I discovered against WhisperzQ may be an example of this - perhaps well known long ago, but brand new to us.

And thank you for your kind words of encouragement! :-D

Fwiffo: If you would like to finish this game by private message, one move per message, with you swapping sides to white and me taking black, I'd be willing to give it a go. We could post the result here afterwards. You can try and stop me winning as black. I'd be surprised if it takes more than 10-15 moves.

Fwiffo: It is a fairly simple win. You would have needed to do two things:

Ensure the king and his fortress companions were sufficiently surrounded to stay in the fortress. This is easy.

Restrict the movement of the outer white pieces until either (a) they have no movement at all, or (b) they have to offer themselves for capture. This takes a few careful moves but is easy.
The win comes either because (A) white has no more available moves, or because (B) he has no moves except one that fatally weakens the fortress.

AbigailII: Good point!

Fwiffo: An example of a true draw without the 4 soldiers in a corner:
http://brainking.com/game/ArchivedGame?g=766596


But this is not a draw!!! It's a sure win for black. As chess masters often (annoyingly) say, "The rest is just technique."

Your opponent was not wrong to resign unless he thought you would throw away the win. Perhaps he was complimenting you by resigning as he assumed that you had the skill to win. Do you think he was wrong to assume that?

Fwiffo: You wrote: "The point is, that two equally strong players should have around the same 'chance' to win. Now, white is already favoured by the rules (60%!!), which is -to say the least- a reason to abolish perpetual raichi."

I would not judge that chance on a single game. The least I would judge it is by a pair of games with swapped colours.

I work with cryptoguru. He is a very strong player with white. He is not so good playing black. He is not alone - he is like many people. Playing white and black are like playing two entirely separate games with separate strategies and even some separate tactics. Playing a successful black strategy takes much more patience than white. It takes more time to master playing black than white for many people because of the attractiveness of forced white wins. This is human nature, it is not a problem with the game itself. Most people here have been playing months not years. When these games were originally played, you'd learn over a lifetime. I think the statistics reflect a lot of learning and hence a lot of bias towards white winning. If you investigated statistics for more mature players and exclude the beginners, I suspect you'd see it even out a lot.

We should allow perpetual raichi because it's part of the game. I don't think it's a goal. It's just an artefact, a side effect. We have no evidence to suggest that it was not a traditional part of the game.

Fwiffo: You wrote: "That the black player could have prevented perpetual raichi doesn't prove white deserves the draw.". No it doesn't prove that white deserves the draw. It proves that black doesn't deserve the win! :-)

If it is really asking too much to ask a black player to stop the draw as well as to prevent white winning, why is it that the very best tablut player on the BrainKing site has so efficiently stopped nearly everyone drawing against him? (I know that he hates the perpetual raichi draw, but that hasn't stopped him being 400 BKR ahead of anyone else AND with less than a 1% draw ratio.)

I know some people play for a draw from the beginning. If you are good enough as black, you will stop them. If you are not good enough, they will draw against you. When you become good enough, you will begin to stop the draws.

I believe that perpetual raichi is only a problem if it can be demonstrated that white can always force it from the start position. To the best of my knowledge that has not yet been done. If it is ever done then the game rules are flawed and will need changing. Until that time, I am happy to play both sides of the game with the current BrainKing rules.

In my humble opinion, if you let white get a draw then it's your own fault. As black you have to play with this in mind. I've looked at one of your games (841545) and I believe you could have won. But instead you took a greedy short-sighted approach instead of patiently encircling your opponent and you let them draw. That's just a statement of fact - it's not meant to be offensive. The important words are you let them draw. It was a draw because you let it happen.

Not accepting a draw would be bad sportmanship. What are you going to do instead - play the game forever? What would that prove? It would say less about the game and more about you.

This so-called "problem" has been raised before. If a solution is required then no single answer has yet been agreed. But perhaps you have a good answer. Do let us know. Meanwhile bear in mind that the "problem" is not a big one. Less than 3% of tablut games are draws.

Since the sudden recent demise of "the Tablut Fellowship" I have created Tablut Reborn. I hope this will become a popular and active fellowship.

cryptoguru: Good stuff Stuart. How about a help screen to explain all the facilities to people?

mikkyT: We are not going to discuss bannings on this board. That's pretty much true as a general rule on any public board but I'm going to stick to it here. If you want to discuss things privately that's a different matter.

Fwiffo: The answer is "Yes", though perhaps one day others will join the pantheon.

mikkyT: Since this is a tournament game and it is still in progress, I would urge readers of this board not to comment on this specific game.

That aside, it is a perfectly legal position arrived at by perfectly legal moves. Black obviously has a huge advantage. If the game ends in a dozen moves time rather than immediately then so what? Such is the nature of Black's normal task in Tablut.

mikkyT: Here's another example:
http://brainking.com/en/ArchivedGame?g=727934
Perpetual raichi in 7 moves and I am the 'culprit'.

Why did I play it? This is in a tournament. I'm playing the site #5 who is 200 BKR points above me. We're level in our tournament section on points though he's played one more game than me. I can take the immediate draw or take the risk of winning but possibly losing. Will a draw satisfy me in those circumstances? You bet. Have our BKRs changed as a result? No.

Could black have stopped this? Possibly. I hope so. Otherwise the Tablut rules are definitely flawed. If black did not want a drawn game, and it was possible to stop it, then he would have (or should have) taken this into account at the outset.

mikkyT: Don't like the game as much as our standard BK variant. With "corner" tablut, you more or less have to allow the king to be involved in capturing black pieces. You can also treat the corner as a white piece and can capture against it, else black just sets up a barricade of two at each corner. But then black sets up a barricade of three diagonally at each corner and he cannot lose (see my draw against WhisperzQ). Not satisfactory in my view. The games become long and boring too, which is not what happens at present.

mikkyT: Well isn't that just the point? If you're playing someone who is 545 BKR stronger than you (as in this case), you are very likely to get hammered! In fact your scoring chances are just 4%. Wouldn't you head straight for draw if you could?

mikkyT: But how many other kinds of circumstances are there?

I checked another forum I know ughaibu has posted to, and his posts have gone there too. Either he has painstakingly deleted them all himself (highly unlikely and there is no reason to do so here), or some entity such as global mod has taken exception to him and deleted all his public forum posts. In the latter case I suspect they have screwed up and/or exceeded their authority since on the Tablut board ughaibu has always been constructive and benign, whereas on, say, Gothic chess, his posts have been more controversial.

Harley: please find out what has happened and get his posts reinstated, if you can.

ughaibu: Yes it does, provided (a) that we don't include the rule 11 that Fwiffo has mentioned, and (b) that white has at least one other piece that can stay mobile outside the black enclosure of the fortress.

The fortress is very easy to set up - easier than perpetual raichi. (For those that don't know what a fortress is, imagine the white king on f6 and a white guard on f5. These cannot be captured with the rules in their current state.) Keeping at least one other mobile white piece is IMHO much harder since he has no raichi threats, and when white runs out of moves he loses.

So in summary, I feel that a fortress draw will very rarely be an option anyway, though of course I yield to those who have much experience than me in evaluating that.

ughaibu: You have a point in that there is no reciprocity, black vs. white, in terms of drawing. And if one sides plays badly right through then losing is a fair result. But in other games, even chess one often has the opportunity to claw back some honour in the result. The oriental game of Go is a perfect example of where you can screw up early on in one part of the board and still have good chances elsewhere and in the game as a whole. The most satisfying games of chess I have ever played have all been hard fought draws so perhaps I appreciate drawn games more than most. Tablut is unbalanced by nature and though it is a shame that black does not, on the face of it, have the resources to force a draw, there can still be times when he cannot capture the king and so will be happy with his side of a perpetual raichi.

furbster: Perpetual raichi is where the white king attacks the edge (=raichi) and black can do nothing but interpose a piece. White then moves the king along the edge to attack a different part of the same edge, and black moves (usually) the piece he moved last time to cover. White moves back to attack the first edge square and so it oscillates. White can keep it up forever and the game can be nothing else but a draw. This can happen when the king can move along a single file or row to attack N edge squares where there are less than N defenders that can be scrambled to interpose.

ughaibu would like to make it part of the rules that perpetual raichi should not mean a draw for white and that instead white should lose. He has the following good reason for this: if (as seems may be possible) white can force perpetual raichi right at the start of the game after only a short sequence of forced moves, then the game as we know it is flawed and not viable. (The same would be the case if either side could force a win from the opening.) I agree, but my take on it is slightly different, namely that I believe there should be drawing resources available in the game, even if they continue to be used in less than 3% of total games as at present. For instance, if white plays really badly in the opening and is fighting for his life then I believe that perpetual raichi should be permitted to save the game.

But how to determine when perpetual raichi would be allowable? I have suggested elsewhere that white might be permitted perpetual raichi as a draw after X moves, where X might be 10 or more. ughaibu thought about that and suggested x = 20. However, he also thinks that it would be tedious to keep track of the move count over the board (if you ever play face to face). I disagree since it is easily possible to have a extra counter that could march down outside one (or two) edge(s) of the board, and which it is black's responsibility to move, and which could easily count 9, 10, 18 or 19 by its position.

An alternative would be to say that perpetual raichi is possible only when white has fewer than Y pieces left, where Y could be 5 (king extra). A further suggestion is that any such draw forces a immediate replay with colours reversed, though this latter suggestion would be no good if we could prove that a draw could be forced from the outset as the match could then go on forever!

So in conclusion, I believe that perpetual raichi should sometimes result in a draw and sometimes in a win for black, but I am yet undecided as to what should determine the threshold between those two states.

Edit: spelling corrections only.

The situation arose because as White I played very badly at the start of the game (and in another similar game). I was still relearning the game after all. I wanted to see if I could salvage a draw rather than be crushed to death.

I discovered that white can build a central fortress for the king. But this is not enough to draw if the rest of his pieces are surrounded and their mobility gradually reduced to zero. White would then lose because he'd either run out of moves or he'd be forced to making a losing move.

So I investigated what configurations of white pieces that could survive on the edge or in the corner and still have moves available due to an inner mobile piece. In the game with WhisperzQ I was able to do it because he didn't spot the possibility.

I agree with ughaibu that this strategem (which may have been previously undiscovered) adds complexity and depth. The central fortress is easy to set up but can prove useless in the face of a patient opponent. But the edge or corner mobile group is very hard to set up, and will be even harder now everyone knows about it!

snigfarp: Fair enough. I would go along with that.

snigfarp: I contest that it would be a more accurate reconstruction if we had the raichi rule, and that we are actually playing a variant now rather than the most authentic version we could be! And why would previous games be invalid?

estanto:

I don't think it complicates the rules at all! When you learnt chess, did you think it a complication that you had to get out of check and that the king wasn't actually captured?

I think it is more satisfying to play this way.

Following my suggestion arguably makes the BrainKing implementation of Tablut more "authentic" as it is more likely to mirror the real game as it was originally played.

What is the point of calling "Raichi!" by the King's side player in the original rules? Obviously to announce that the King can 'see' the edge of the board. This is equivalent to "Check!" in Chess.

But what's the point? The point must be that the defensive player is alerted to this possibility and must block it. At chess, the defensive must get out of check if he can. At BrainKing we don't play Raichi and if a defender fails to spot that the king can see the edge, the king's side will win on the next move. I think games that end like that are rather unsatisfying.

Therefore I propose the following small change to the BrainKing rules and implementation of Tablut:

If the king can see one edge, it is Raichi and the defender must try to block it if he can. The interface should not allow a non-blocking move. If he has no blocking moves then king's side wins straightaway;

If the king can see two edges, it is Tuichu and the king's side wins straightaway.

I think that would be much better. Please let me know your views.