Ratings & Comments
Some time ago I tried using the game code wizard to make a preset for a large (18x18) game using a few crooked sliding pieces. Now I notice that one of them, the Hornet from Seenschach, works in the applet but not in the generated code. The way I defined it was Wafs(alar)7Wafsal(aral)7WzB. Part of the move works, but the zB seems to be ignored, even though the generated code appears to include those moves as well.
This looks interesting, but I wonder how well it works allowing the maggots to move backwards.
That describes the 'policy head' of the NN, which is used to bias the move choice (which is otherwise based on the number of visits of the move and that of the total for the node, and the move scores) when walking the tree from root to leaf for finding the next leaf to expand. But my understanding was that when the leaf is chosen and expanded, all daughters should receive a score from the 'evaluation head' of the NN in the position after the move, rather than just inheriting their policy weight from the position before the move. These scores are then back-propagated towards the root, by including them in the average score of all nodes in the path to the expanded leaf.
I do not understand "tgat". You probably meant that.
Indeed. Hard to avoid typos on these virtual keyboards of Android devices... I corrected it.
I'm not sure I understand what you say, HG!
What is not understandable? Typo! What else.
By the way, details about programming are not clear for most people. How to deal with it?
I did not meant the typo!
By the way, details about programming are not clear for most people. How to deal with it?
I don't think there is any "fix" to this issue. I am not sure there is any issue at all. Some conversations are going to involve things other people don't understand. That said, the talkchess forums are the usual place for these kinds of discussions, but I am happy to have some discussion here as well. Some people who are not chess programmers may still be interested in whether the new neural-network techniques being applied to orthodox chess can be applied to chess variants.
That sounds plausible.
I have somewhat of a dilemma concerning the move notation for castling in the Interactive Diagram. Normally King-side castling is O-O, Queen-side castling is O-O-O, both for white and black. This then corresponds to short and long castling, respectively.
But what if the Kings start closer to the a-file? Would it still make sense to keep calling the a-side the Queen side, and use O-O-O for that castling. I know that the official notation for Chess960 does this, but that is really another case, because the King there can start anywhere, but at least ends on the c-file, like in orthodox Q-side castling. So it is indeed like a long castling, only with messed-up initial position because of the shuffling.
But what if the King in a-side castling ended on the b-file. Does it still deserve to be written as O-O-O? I encountered this problem in Elven Chess, which is unusual in that it has rotation symmetry rather than reflection symmetry in the initial setup; usually variants that have that do not have castling, but Elven Chess does. The white King starts on the f-file, and moves 3 spaces to i1 on the 10-wide board. So it would be normal to call that O-O, and the castling to c1 O-O-O.
But now what for black? His King starts on e10, and castling would bring it to b10 or h10. I would be inclined to call the castling to b10 O-O now, not O-O-O.
Any ideas what we should elevate to standard here?
I don't think the O-O, O-O-O thing scales very well. What about games with an odd number of files and the King starts in the middle?
You could just use O-O for everything along with the king's destination square, like O-O b1. If you want to preserve O-O-O, it could be understood as referring to the left half of the board from White's perspective, while O-O refers to the right half.
You could just use O-O for everything along with the king's destination square, like O-O b1
You could. But, at that point, I would ask what value the "O-O" is bringing. Wouldn't the more standard "e1b1" also accomplish that? Although, I guess this would address the issue with (e.g., Wildebeest Chess) where the King can move a single space and still castle. Perhaps if either "O-O" or "O-O-O" is followed by the notation of a square, it would mean castling to that square (and there would be no difference between "O-O" and "O-O-O")
To solve the ambiguity problem in 1-step castling (or in other castlings where the castling piece can also move multiple squares on its own, such as that silly 'guarding' of the Queen in Enhanced Omega Chess), the Diagram uses the tilde instead of the hyphen (or nothing) as connecting sign for indicating castlings, like K~d1. But I like to stick to conventional SAN where this is possible.
Well, I'd like to point out the grand apothecary chess games where you may castle in four ways with 2 different pieces!
I think the most general solution is to point out both moves and then add a simple "c" for example.
Hi H.G.:
I didn't change anything for the longest time in the diagrams on the Sac Chess rules page. I assumed the chancellors disappearing might have resulted from something Fergus (or someone else) changed on CVP website, Game Courier or the Diagram Designer. Right now I haven't tried to edit a rules page myself for so long (if that's what needs to be done) that I'm afraid I might make things worse.
I guess the diagrams are generated 'on the fly'. So probably the image file for the chancellor was renamed.
And now that you are here: I have been commenting a Sac Chess game for Jean-Louis' new book, and I noticed that after castling the King is really still very unsafe. Because it is still directly behind the Pawn shield. So castling really seems a bad idea. It seems much better to quickly move out an Amazon to f3/f8 or e2/e9, and then hide the King away on g1/g10. The Rook is then not trapped, and can get out once the minors are developed, and the Sailor can be moved to i2/i9 to protect the Pawn Shield. The Judge, Archbishop and Chancelor can then get out over h1/h10, g2/g9 and h2/h2, respectively.Leaving the King behind an enormous 'wall of power'.
Does that make sense?
I'd hoped that castling would often come in handy, but there have been a couple of games, so far, where my opponents have dropped their K back a rank against me (likely after an Amazon move to third or second rank - one possible drawback might be if said Amazon ever has to retreat somewhere with loss of time). Castling kingside followed by shifting my K sideways one more square is something I often do, though maybe out of habit from chess (somewhat unrelated, Play Tester recently championed the idea of quickly charging the pawn in front of his Sailor on the kingside, even before either side may have castled).
The more space you have control of on a side of the board, the safer it is to castle there, I suppose. If Black plays a French Defence analogue, for example, then kingside castling seems nice enough for White. Not only that, but if a centre file is about to get opened early, castling may be a good idea. Sac Chess is still relatively unexplored, of course. I thought I'd sensed some unexpected defects to its design (compared to chess, at least) since inventing it, but at least it's being played more often again lately (though I'm always one of the players). Bishop(s) (and later Missionaries) flying out to the edge of the board, especially of the queenside, for example (the Sailor pawn charge may be another).
I'd also hoped when designing the game that the Judges (Centaurs) on the wings would help to guard a castled K for a long time. Castling queenside seems like it's usually quite unsafe, even compared to chess. The K is still likely unsafe in the middle for a couple of moves at the least. I'd secretly hoped to be the first one to castle queenside in a Sac Chess Game Courier game, but Fergus beat me to it when playing someone else. It's also easier to discourage or prevent castling queenside in the first place than in chess, it seems.
I would like to see similar discussions around 'AC'.
Testing the game first with actual play, say using a preset on this website's Game Courier, would give players and viewers much clearer ideas about how any complex strategies might unfold in typical games.
Because the game is not played on a board that uses plain graphics (like a chess board), you (or someone else) may need the help of a CVP site editor or the webmaster to get the graphics of the board onto this website, for subsequent use on Game Courier when you (or someone else) will go about making a preset.
The ZoG file doesn't seem to work, giving error 'The following bitmap couldn't be loaded: "images\boards\rwb10x10bmp" '
If you could advise on a fix, I'd love to play this
Would it be possible to add a Wildebeest (=Knight+Camel) to this fairychess? Thank you
This is not a game in itself, but a book about a game. I found no other way to create a page about it, so I hope this works out. Let me know otherwise.
You can play this with the latest ChessV release candidate: ChessV 2.3 RC2
Just unzip and run the EXE.
28 comments displayed
Permalink to the exact comments currently displayed.
For Diagram Designer it is same? I uploaded my files in chessvariants , but what I should do to upload in Diagram Designer?