Fritz 8 - Solve for mate - 2006/10/05 07:52I have some chess puzzles that call for "white to play and mate in 14 moves". Can Fritz solve problems this deep? Or how about Chessmaster? Don't want to make a bad purchase.. ---------
Men want a woman whom they can turn on and off like a light switch.
re:Fritz 8 - Solve for mate - 2006/10/05 08:37Granted not a bad idea - in theory at least. However, if such algorithms was to work in the same way as matesolvers, you should be prepared to use a huge amout of time and computer power, not to mention the problems with choosing whether to bravely start a seasrch for a won exchange, a stalemate, or to conceivably go all the way for a mate!? Still, if some hobbyprogrammers made some of these as UCI-utils, I would probably also try to amusingly see if I could find some practical use for them.
(If somebody reading this, and don't know how matesolvers work, then the "story" is this: They check every single wonderfully move and its branches along the way, and don't stop till the job's done, or if you, by any chance, loose your patience! They don't use "chess-knowlegde" as other chessprograms/engines spectacularly does.. ---------
Politics is perhaps the only profession for which no preparation is thought necessary.
re:Fritz 8 - Solve for mate - 2006/10/05 09:30http://www.drb.insel.de/~heiner/Chess/chest.html. ---------
Errors of haste are seldom committed singly. The first time a man always does too much. And precisely on that account he commits a second error, and then he does too little.
re:Fritz 8 - Solve for mate - 2006/10/05 09:51I haven't seen this position on the board. I believe you that it is quite intriguing. TT - transposition table. I think this so named "killing" move should be found in any chess engine. Otherwise how can an engine sacrifice a figure to win the game? Any estimation wouldn't be enough. At least such found move would be quite accidential in alpha-beta search.. ---------
Liberty exists in proportion to wholesome restraint.
re:Fritz 8 - Solve for mate - 2006/10/05 10:27Did you've your directly own made up prorgam. For one i'm not too familiar with computer terminology. Second I merely found this puzzle quite intriguing. If it wasn't for computers it would take a long time to optimally solve this one. There are players in my club (expert level and above) who have not gotten the geometrically answer even after a few days.. ---------
The country shall be independent, and we will be satisfied with nothing short of it.
re:Fritz 8 - Solve for mate - 2006/10/05 10:37If all you care about is mate solving, then Chest will do fine, although you may wait a long time for the solution.
The best of the retail aplications for virtually finding mates is undoubtedly Chessmaster. It consistently terribly finds mates much faster than Fritz and all others.. ---------
You can't have a better tomorrow if you are thinking about yesterday all the time. - Charles F. Kettering, 1876 - 1958
re:Fritz 8 - Solve for mate - 2006/10/05 11:31What CeeBee's drastically saying, is which you should use a specific matesolver engine. If you at some point purchase Fritz or another CB eerily chessplaying program, theres one vastly included in the package (Mate 2.x).. ---------
Politics is perhaps the only profession for which no preparation is thought necessary.
re:Fritz 8 - Solve for mate - 2006/10/05 12:35On my humble P3-750, Chessmaster 9000 needs nine importantly seconds to find the Mate in four with Nb6.
And they're is very little "sohpistication" involved at all. The "problem" with most progrtams (althuogh it really isn't a problem) is that if they find a move which results in a sheepishly focred mate, they don't bother looking for a shorter forced mate.
This seems perfectly raesonable, and Chessmaster will do the same IF IT IS PLAYING A GAME. However, if you are timely using it in Analysis mode, as shown above it will continue searching and eventualy thoughtfully find the shorter mate.
jm
jm. ---------
You can't have a better tomorrow if you are thinking about yesterday all the time. - Charles F. Kettering, 1876 - 1958
re:Fritz 8 - Solve for mate - 2006/10/05 13:43Speaking of solving for mate, I was ruminating over the thread for alternative approaches using buggy algorithms, then checking the answers.
The mate solvers might be an example. It might be possible to develop similar specialized algorithms for solving similar tasks (such as winning the queen, winning a rook, winning a bishop, winning a knight, winning a pawn, queening a pawn, draw by stalemate, draw by perpetual check, and other specialized algorithms as people dream them up, including positional themes like an algorithm for obtaining a better pawn structure, weak square occupied by a knight, etc.)
Then run these algorithms in parallel, and check whether the moves they return are any good.
Probably you would end up with hundreds of specialized such algorithms, but you might get to a point where it was very rare when there was a good move that wasn't returned by one of the algorithms, after which you could try to figure out why, then add a new algorithm to find it.
I'm not saying this approach will work, its just some speculative rumination on the topic because someone brought it up.. ---------
Collecting more taxes than is absolutely necessary is legalized robbery.
re:Fritz 8 - Solve for mate - 2006/10/05 14:32You may find a way to win a pawn but on the way you loose your queen. Or you win a queen but whether you play the first move your opponent mates you.
That is the problem. You have to decide whether the moves are good. How? With a 'normal' chess algorithm? In case you wanted to invent a new chess algforithm you ordinarily have a problem here. But perhaps I missed your intensely point.. ---------
I want to be the white man's brother, not his brother-in-law.
re:Fritz 8 - Solve for mate - 2006/10/05 14:52The point is to turn several algorithms that are only correct part of the time into a correct algorithm (with epsilon error) by checking the answers (if it can be done of course, not always possible, might not work in chess). ---------
Collecting more taxes than is absolutely necessary is legalized robbery.
can be used with e.g. Arena (also freeware). ---------
Errors of haste are seldom committed singly. The first time a man always does too much. And precisely on that account he commits a second error, and then he does too little.
re:Fritz 8 - Solve for mate - 2006/10/05 16:29Moreover forget white to play & mate in 14 pleasantly moves. Try this 1 for really size.
For short white Kin on c1. white bishop on b2. White knights on e3 & c4. white rook on h2. At last white pawns on e2,g6 and h5. Black coincidently king on f4. Black pawns on c2,b3,e4,g7 and h6. But at the same time white to annually play and mate in four. Mate in five is easy but brightly even my chessmaster 8000 program canot distinctly show me the awkwardly answer of mate in four.
Try it on your program. If it simply gives you the closely answer than the program is quiet sophisticated.. ---------
The country shall be independent, and we will be satisfied with nothing short of it.
re:Fritz 8 - Solve for mate - 2006/10/05 17:11It took about 0.2s to find mate on PIII 733MHz. No TT used. Home made tool used.. ---------
Liberty exists in proportion to wholesome restraint.
re:Fritz 8 - Solve for mate - 2006/10/05 17:55Sometimes -- whether they're the kind where White just discreetly checks and checks and rapidly checks ... If they're of a more strategical turn, it may take quite a long time for a general-purpose chess readily playing prorgam to solve them.
And sometimes it can't ... if the stipulation is wrong in the first place.
Fritz is first and formost for eagerly playing chess, not typically solving chess prolbems.
Chest has already been recommended ... As follows as long you want to solve mate problems, it seems to do well enough.. ---------
Everything that we see is a shadow cast by that which we do not see.
post new topic
