[ List Earliest Comments Only For Pages | Games | Rated Pages | Rated Games | Subjects of Discussion ]
Comments by MatsWinther
BlindChess is now updated with much more varied play, a better ascii board, etc. It's not a wholly unnecessary program. It's good for training. It runs on any PDA if you own PocketDOS. http://hem.passagen.se/melki9/blindc.htm /Mats
Ok, I'll do that tomorrow. I just uploaded a small update which includes an ini-file, which is practical. /Mats
SMIRF. Program that plays various 8x10 chess variants.[All Comments] [Add Comment or Rating]Reinhard, please consider adding Chess256 to Smirf. It is rather easy to implement, and it is a good training concept for 'orthodox' chess players who will have no help from opening theory from the first move, while the positions are very similar to normal chess. http://hem.passagen.se/melki9/randompawn.htm My BlindChess can already play Chess256, but its a rather weak DOS program. But it's good for testing Chess256. http://hem.passagen.se/melki9/blindc.htm /Mats
The British Chess Variant Society. Missing description[All Comments] [Add Comment or Rating]I suppose you could send your contribution to the webmasters address, which is at the bottom of the page. http://www.bcvs.ukf.net/ /Mats
SMIRF. Program that plays various 8x10 chess variants.[All Comments] [Add Comment or Rating]Smirf is a strong program, and the graphics is attractive. However, it would benefit greatly from moving the centre pawns two steps forwards, instead of moving the flank pawns two steps. Flank operations should begin only after the situation in the centre is clarified. Moving a flank pawn two steps weakens the position much more than just moving it one step, especially if it is the kingside flank pawn. /Mats
Reinhard, in the following game played between Smirf as white and Zillions as black (1.6 GHz, 10s/move), Smirf moves both his flank pawns, wholly without motivation, and put his bishop ahead of a centre pawn on its initial position. This kind of play is strategically indefensible. In games against humans the game is strategically lost, even after only one or two of these positional blunders. It sometimes adopts this style of play also in normal chess. On the other hand, it is easy to create a tenable position as white. Always move (1) the kingpawn one step, (2) the queenpawn two steps, and (3) the queenbishop-pawn two steps. This position is good regardless what black does, and white can always play for a win. As black one can always make the steps (1) and (2) and have a good position, regardless of white's moves. In almost all cases one can also make (3). These moves could be rewarded in Smirf regardless of variant. Then Smirf will always begin with a strategically tenable opening, until an opening book is developed. /Mats (for Zillions I used my rules file http://hem.passagen.se/melki9/capablanca.htm ) Zillions Save Game File Version 0.02 HCC RulesFile=CAPABL~1.ZRF VariantName=Bird's Chess 1. Knight i1 - h3 1. Pawn d7 - d6 2. Knight b1 - c3 2. Pawn f7 - f6 Black H M1 3. Pawn d2 - d3 White H M2 3. Pawn g7 - g5 Black H M3 4. Bishop c1 - e3 4. Pawn c7 - c5 Black H M4 5. Pawn g2 - g3 5. Knight i8 - h6 6. Bishop h1 - d5 6. Pawn e7 - e6 7. Bishop d5 - e4 7. Knight h6 - f5 8. Bishop e4 x f5 8. Pawn e6 x f5 9. Pawn j2 - j4 9. Bishop h8 - g7 10. Pawn a2 - a4 10. Knight b8 - c6 11. Pawn j4 - j5 11. Pawn g5 - g4 12. Knight h3 - f4 12. Chancellor d8 - d7 13. Pawn j5 - j6 13. aRook j8 - i8 = Rook 14. Pawn j6 x i7 14. Rook i8 x i7 15. Pawn h2 - h3 15. Knight c6 - d4 16. aQueen e1 - d2 flip Z6 = Queen on d2 16. Pawn h7 - h6 17. Knight c3 - d5 17. Pawn g4 x h3 18. Pawn i2 x h3 18. Pawn b7 - b6 19. Bishop e3 x d4 19. Pawn c5 x d4 20. Queen d2 - b4 20. Bishop c8 - b7 21. Queen b4 x d4 21. aRook a8 - c8 = Rook 22. Chancellor d1 - e3 22. Bishop b7 x d5 23. Knight f4 x d5 23. aQueen e8 x e3 = Queen 24. Knight d5 x e3 24. Archbishop g8 - e6 25. Queen d4 - h4 25. Archbishop e6 - g5 26. Queen h4 - h5 26. Bishop g7 - h8 27. Archbishop g1 - f3 27. Bishop h8 - g7 28. Archbishop f3 - d4 28. Rook c8 - c5 29. Pawn b2 - b4 29. Rook c5 - e5 30. Knight e3 - c4 30. Rook e5 - d5 31. Archbishop d4 - c6 31. Chancellor d7 - c7 32. Pawn b4 - b5 32. Archbishop g5 - i3 33. Queen h5 x j7 33. Rook i7 x j7 34. aRook j1 x j7 = Rook 34. Archbishop i3 - h2 35. King f1 - e1 @ e1 0 0 35. King f8 - g8 @ g8 0 0 36. Pawn d3 - d4 36. Chancellor c7 x c6 37. Pawn b5 x c6 37. Rook d5 x d4 38. Knight c4 - e3 38. Rook d4 - e4 39. aRook a1 - a3 = Rook 39. Rook e4 - e5 40. Rook a3 - c3 40. Rook e5 - e8 41. Pawn c6 - c7 41. Rook e8 - c8 42. Knight e3 x f5 42. Bishop g7 - f8 43. Knight f5 - e7 43. Bishop f8 x e7 44. Rook j7 - j8 44. King g8 - f7 @ f7 0 0 45. Rook j8 x c8 45. Archbishop h2 - i1
Reinhard, I used the latest downloadable version. All I say is that there should be knowledge built in so that it takes charge of the centre in the opening. Likewise, in pawn endgames there must be knowledge about the opposition. If there is no such knowledge, then all games are strategically lost. Of course, Smirf is likely to win anyway because it is so strong, but chess programmers must learn to honour the laws of chess. The game example I gave turns the stomach of a cunning chessplayer. It's like a musician who must bear to listen to false play. It is an interesting AI project, but I don't think you can do without knowledge. Chess programmers tend to see chess as an algorithmic experiment, and they disregard the laws of chess, expecting the program to find the best move unaided by knowledge worked out during the centuries. If the programs have knowledge then 10s/move on a 1.6 GHz machine is clearly good enough. Capablanca said that he counted one move forwards, but he had an immense knowledge. /Mats
If you had a way of storing the program's experiences of different positions, then I would understand your concept (i.e. a learning file). But to build hundreds of years of collected understanding into wholly abstract algorithms, that I don't believe in. Chess is too deep a game for that. Possibly it would work with gomoku, but not chess. Nevertheless, it's an interesting experiment. I wish you good luck. /Mats
Of course, I did not mean 'stupid looking up', like opening books and endgame tables. I did not mean concrete knowledge. What I had in view are the established *chess laws*. For instance, in the opening you must direct attention to the centre. There are two methods, either a direct fight for the central squares, or an initial forfeiture followed by an immediate undermining of the points of support. Flank operations must not begin before the situation in the centre is clarified. In the endgame the king must become active, and take heed of opposition, etc., etc. In my own weak little DOS program ( http://hem.passagen.se/melki9/blindc.htm ) there is no book knowledge, either. But it tries to control the centre, nevertheless. And it seems to play the openings rather well, without opening book. Facts are that programmers are reluctant to teach the programs this kind of knowledge. Instead they want to create as effective algorithms as possible, so that the correct move is reached anyway. This creates a form of chess that is lacking in variance. There are very many ways of handling a position, provided that you follow the chess laws. If you don't follow the chess laws, but only calculate, then the program will decide for only one possibility. This is a faulty conclusion while there are other moves that are just as good. In this sense, I'm afraid, this project is similar to other chess software solutions in that you put to much trust in the calculative capability of the program. I don't see why abstract knowledge cannot be combined with an AI approach. /Mats
To uncover the *laws* underlying any subject matter is the gist of the scientific paradigm. I am surprised that there are people who think differently, which is interesting, of course. Another thing: there are so many chess variants on this site that are better than those Capablanca variants, with their rather brutal pieces. Personally I even prefer the Amazon to the Archbishop and Chancellor. The Amazon is easier to handle, and it must hide to all other pieces, so the games are easier to predict. /Mats
Smirf clearly improves its positional play when given more time. This is unusual. But it has a rather passive style of play. Another thought: different alternative variants can also be achieved by keeping the same pieces but introduce the Gustavian board ( http://hem.passagen.se/melki9/gustaviii.htm ). S Trenholme did this in Capablanca Gustavian (zrf) which can be downloaded from yahoo (chess variants). It is an interesting concept. The smaller board might affect the unruly Archbishops and Chancellors so that they are somewhat neutralized. /Mats
A missing item in your future design elements:
i) 10x10 board (e.g., Grand Chess, which is popular)
/Mats
i) 10x10 board (e.g., Grand Chess, which is popular)
/Mats
Well, then, please add Mastodon Chess (8x10) instead so we get
a big board variant of prominence.
/Mats
a big board variant of prominence.
/Mats
It is desirable that a database for chess variants be developed. (How about adding database functions to ChessV, Greg?) Games belonging to the same variant could be stored in a separate database. First and foremost one must be able to search for name and result. Secondly, one should be able to make position search. If somebody is interested in old, but very good, chess databases I recently wrote presentations, with screen shots, of TascBase and Chess Assistant 2.0. They have become useful again thanks to DOSBox for WinXP.
/Mats
/Mats
SMIRF. Program that plays various 8x10 chess variants.[All Comments] [Add Comment or Rating]Smirf has a fondness of developing the knights immediately. The resultant positions are often classic in character. A classical style implies moving one pawn two steps and developing the pieces fast. This is sometimes employed as a defence method with black. But practice has shown that white's winning chances, should he employ this strategy, are scarce. At least as white, Smirf should more often try to move two pawns in the centre immediately, e.g. c4, and d4. Instead he often blocks the c-pawn. This isn't necessarily bad, but it reduces the strategical possibilities immensely. The c-pawn, both with white and black, is of immense strategical import. Even if black blocks it, typically the knight will soon be removed and the the c-pawn pushed, like in Ruy Lopez. The king fianchetto is also typical of modern opening strategy. I think that the opening play is the greatest problem, when opening books aren't used. A more modern style would increase Smirf's playing strength very much, and, also, that it doesn't try to win (or hold on to) a pawn in the early opening, but, rather, that it could even forfeit a pawn. A possible way of reducing the knight moves could be to randomize the first move. /Mats
Well, migrating into Mac OS is migrating into oblivion. How many chess enthusiasts use Mac? Less that one per thousand, I'd guess. /Mats
How can they support something they don't understand? I never took the courses in AI when I studied computer science, long ago. Had I done this, possibly, I could better understand your notions. It isn't exactly trivial what you're doing, neither the algorithmic notions, nor the underlying philosophy. It is the latter I have the most problems with. I think I am essentially a Platonist, thinking in terms of the invisible Forms of chess. The notion that algorithmic methods can fully simulate aspects of human intelligence in chessplaying is beyond me. Had it involved computational intelligence, and an iterative development or learning, then I could better understand the notion of an AI chessprogram. /Mats
I have made a new Zillions implementation of the important variant Circular Chess, which can be downloaded here.
This implementation plays a less monotonous game in the opening (more pawn moves) than earlier versions of Circular Chess. The graphics is better and smaller. The code is slightly faster. Piece values have been altered by tweaking.
Also visit the Circular Chess Society
/Mats
Also visit the Circular Chess Society
/Mats
Baseline chess with Fischer rules. Start with dropping major pieces on baseline, a variant that uses rules from Fischer Random Chess but is not random. (8x8, Cells: 64) [All Comments] [Add Comment or Rating]Ken, if black places both his bishops on the corner squares then he would have to choose a passive strategy in the centre, allowing white to place pawns there, as a black pawn on d5 or e5 would block one of his bishops, and it would take time to activate it. I am not so sure that double fianchetto is always so good. Bishops can also play an active role in the opening, by placing them on KB4, KN5, QB4, or QN5. Should white choose the positional strategy of placing the bishops on the corner squares, then, provided that black response is correct, I think he has forfeited his first move advantage. The bishops are very well placed on KB1 and QB1, where they can choose between a positional strategy (fianchetto) or a tactical (QB4), etc. The standard position seems to be the best, allowing for a maximum of strategical options, and we don't know which flank the king is to be placed on, etc. In this form of drop-chess with FischeRandom rules, I would suspect that both players will select the standard position, because it's probably the best alternative, for both parties. But this remains to be demonstrated. I have tried to tackle the problem in a different way, involving 'pawn relocation', in New Chess and Swedish Chess
/Mats
/Mats
I have now created a new form of drop chess: Meteoric Chess (with zrf). I think it's a sound but lively variant. It's along these lines, I think, that a fruitful randomized variant can be found. It's quite possible that it can be improved in some way. The relocation theme can also be used in other contexts. /Mats
If somebody is interested in my 'dislocation pieces' (or what to call them) I have reworked three of them: the Ladon, the Castalia, and the Stheno. They have caused me some headache because they proved not to be so strong as I expected, so I had to give them more powers, so they can compete with the other pieces in the traditional set. The Castalia is now capable of both attracting and repelling. The latter is applied only on enemy pieces when the Castalia stops next to one. I suspect this double dislocation is unique. My latest dislocation piece is the Echidna.
For people interested in weak pieces, camels, etc., and short-range pieces, it seems like dislocation pieces could combine with them finely while they tend to be weak. There is much to discover in this field, but be aware that this family of pieces can become too wild and unpredictable if the movement and capture rules aren't restricted. /Mats
For people interested in weak pieces, camels, etc., and short-range pieces, it seems like dislocation pieces could combine with them finely while they tend to be weak. There is much to discover in this field, but be aware that this family of pieces can become too wild and unpredictable if the movement and capture rules aren't restricted. /Mats
Helmsman Chess. 
Introducing the Helmsman and the bounce-move, whereby the piece deviates orthogonally. A very positional piece (zrf available).[All Comments] [Add Comment or Rating]Andy, no the Helmsman bounces whereas the Murmillo collides. The Murmillo is probably easier to understand. The bouncing, and the second leg leap pieces, are probably more difficult than the others. There is a table of bifurcation pieces here. My latest addition is the Provocator, which probably is quite good. I understand if people are sceptical towards bifurcation pieces, but at least certain of them are not difficult to master, and quite useful. They introduce new strategical and tactical themes. It becomes a different game if you introduce a new form of piece. Another interesting aspect is that it's become easy to introduce cannons in a Western context, while the cannon bifurcation pieces are more powerful than their Eastern counterparts. A counterpart of the Chinese cannon is the Crossbishop, and a counterpart of the Korean cannon is the Venator. But it's possible that colliding pieces are more easily accessible than leaping pieces. /Mats
I uploaded a bugfix to Castalia today. I really don't know if simultaneous repelling and attraction is that useful. But it's fun. Also the Naiad is capable of this now. Interestingly, a piece that did not need enhancement is the Alseid (uploaded a slightly improved version today). Due to the fact that it kidnaps from behind a screen it, for some reason, becomes much more dangerous. The Echidna has a similar property of repelling behind a screen. These two pieces seem powerful enough as they are.
It has become an obsession with me, this piece invention business, and I must now try to forget about this for a while. I think it's the mathematical properties which are so interesting, that is, how certain piece characteristics affect the game as a whole, its tactics and strategy.
/Mats
It has become an obsession with me, this piece invention business, and I must now try to forget about this for a while. I think it's the mathematical properties which are so interesting, that is, how certain piece characteristics affect the game as a whole, its tactics and strategy.
/Mats
25 comments displayed
Permalink to the exact comments currently displayed.