![]() |
| If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|||||||
| Tags: against, chess, further, getclub, improved, rybka, seen |
|
|
|
Thread Tools | Display Modes |
|
#11
|
|||
|
|||
|
Sanny wrote:
On Apr 11, 11:59 pm, Bjoern wrote: Sanny wrote: The game was much improved after this modification. Maybe the next improvement could be that it follows the rules? It still can't do that properly. E.g. the three-fold repetition rule or the 50-move rule? Yes, it do not have 50 move rule. Incase it has insufficient material it will however automatically give Draw. Only when it has Rook + then it will try to win the game. I saw your game where Beginner beat you after a very difficult match. No, it didn't beat me. It just declared the game won for itself for no reason (except for being up on material). Looking at the game score it clearly did not know how to win. However it follows three fold repitition draw. You cann see today a game played between Help Bot and Beginner which resulted in draw. No, it does not follow the rule, look at my game against beginner, I repeated a position 3 or 4 times and did not give me an possibility to claim a draw. So it lost only once. Now Help Bot will see a lot of games where Beginner will win against him or atleast have a Drawn game. No, it does not win, it cheats and doesn't follow the rules. |
| Ads |
|
#12
|
|||
|
|||
|
On Apr 12, 4:48 pm, Bjoern wrote:
I saw your game where Beginner beat you after a very difficult match. No, it didn't beat me. It just declared the game won for itself for no reason (except for being up on material). Looking at the game score it clearly did not know how to win. No, it does not win, it cheats and doesn't follow the rules. I played over that game and I believe there may well have been several repetitions of position; I was surprised when the GetClub program finally varied and went on to actually make headway. At the end, there was no explanation of the result-- did Bjoem resign (or did the program detect a mate-in-twenty and forfeit him)? My guess is he got fed up after it repeated the same position over and over, and clicked the "resigns" button several moves later when it became apparent that the program *did* know how to win; it was in fact forcing his King to the edge of the board, and using all of its men for the final assault (how odd!). The GetClub program did work more-or-less correctly in two recent draws against me; but for all we know, it may have only done so because it thought it was "saving" an inferior position by granting the draw. When ahead in material, the program seems far more reluctant to work properly in granting draws. I have "resigned" many of these, in order to just be able to play more games and not be stuck in repetition-hell for all eternity. In another recent game, I noticed that it still moves its own King into check on the final move, rather than try and delay checkmate as long as possible like most programs. But there definitely seems to be a big improvement in its strength of play. I will from now on be looking to trade the Queens off, to avoid getting "thumped" the way I was in the recent game where I saw no significant problem, but in fact walked into a very neat mating net. Yowie-- I was actually outplayed by the GetClub program?!!! LOL -- help bot |
|
#13
|
|||
|
|||
|
No, it didn't beat me. It just declared the game won for itself for no
reason (except for being up on material). Looking at the game score it clearly did not know how to win. No, it does not win, it cheats and doesn't follow the rules. I saw the Recorded game below http://www.getclub.com/playgame.php?...183&game=Chess The Beginner declared win as it saw a Mate in 4. Last position was White: K-d1 Black: K-f4, R-g2, N-e4 And it was Black to move. Black found Mate in 4 So declared itself a win. 1. ... Ke3 2. Kc1 Kd3 3. Kb1 Nc3+ 4. Kc1 Rc2# (Mate in 4) I analyzed with another computer and found a Mate in 4. Thats the reasion GetClub gave itself a Win. I am happy to see GetClub was able to see Mate in 4. Play Chess at: http://www.GetClub.com/Chess.html * At the end, there was no explanation of the result-- did Bjoem resign (or did the program detect a mate-in-twenty and forfeit him)? *My guess is he got fed up after it repeated the same position over and over, and clicked the "resigns" button several moves later when it became apparent that the program *did* know how to win; it was in fact forcing his King to the edge of the board, and using all of its men for the final assault (how odd!). It was Mate in 4 as per analysis by other computer. 1. ... Ke3 2. Kc1 Kd3 3. Kb1 Nc3+ 4. Kc1 Rc2# Play Chess at: http://www.GetClub.com/Chess.html * TheGetClubprogram did work more-or-less correctly in two recent draws against me; but for all we know, it may have only done so because it thought it was "saving" an inferior position by granting the draw. *When ahead Yes, I have seen many games wwhere Draw has come by repitition of moves. When GetClub is at advantage, I have told it to change its move So that repetition with draw do not happen. So After 2 repetition it changes its move to avoid draw. Play Chess at: http://www.GetClub.com/Chess.html * In another recent game, I noticed that it still moves its own King into check on the final move, rather than try and delay checkmate as long as possible like most programs. Yes, When it finds a Mate in 3/ 4. It feels all moves are equally bad. And it gives equal points to all moves and Since king has to die it sometimes plays King into Check. But when it plays such move It has already Resigned. User only sees such moves when looking at recorded Games. Player never sees that move being made while playing. Play Chess at: http://www.GetClub.com/Chess.html * But there definitely seems to be a big improvement in its strength of play. *I will from now on be looking to trade the Queens off, to avoid getting "thumped" the way I was in the recent game where I saw no significant problem, but in fact walked into a very neat mating net. *Yowie-- I was actually outplayed by theGetClubprogram?!!! *LOL Now if you think harder you can win Beginner with a Pawn odds. Easy Level can win you with a Pawn Odd. meaning if you have a extra pawn then your game will be equal to that of Easy Level. Here is the new Rating I give to all Beginner: 1900 (5-10 sec/ move) Easy: 2100 (20-30 sec/ move) Normal: 2300 (1-2 min/ move) Master: 2500 (5-10 min/ move) Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html |
|
#14
|
|||
|
|||
|
Sanny wrote:
I saw the Recorded game below http://www.getclub.com/playgame.php?...183&game=Chess The Beginner declared win as it saw a Mate in 4. It shouldn't do that, it should play out the mate, unless I resign. TheGetClubprogram did work more-or-less correctly in two recent draws against me; but for all we know, it may have only done so because it thought it was "saving" an inferior position by granting the draw. When ahead Yes, I have seen many games wwhere Draw has come by repitition of moves. When GetClub is at advantage, I have told it to change its move So that repetition with draw do not happen. So After 2 repetition it changes its move to avoid draw. It did not do that and the game in question was a draw according to the repetition rule: on move 86 (reptition number 1 is 87.Kf2, then 89.Kf2, then 91.Kf2, when making this move I should have had the option to claim a draw, then further reptitions: 93.Kf2, 95.Kf2, 97.Kf2, 99.Kf2, only then did the computer deviate): 8/8/8/4k3/8/4K1r1/7n/8 w (white king on e3, black king on e5, black knight on h2, black rook checks on g3) 86. Ke3 Rg3+ 87. Kf2 Rg8 88. Ke3 Rg3+ 89. Kf2 Rg7 90. Ke3 Rg3+ 91. Kf2 Rg8 92. Ke3 Rg3+ 93. Kf2 Rg7 94. Ke3 Rg3+ 95. Kf2 Rg8 96. Ke3 Rg3+ 97. Kf2 Rg7 98. Ke3 Rg3+ 99. Kf2 Rg8 100. Ke3 Ng4+ 101. Kf3 Kf6 102. Kf4 Nh6 103. Ke4 Rd8 104. Kf4 Rd6 105. Ke4 Nf7 106. Kf4 Kg6 107. Ke4 Ng5+ 108. Kf4 Rd5 109. Ke3 Kf5 110. Kf2 Rd3 111. Ke2 Rg3 112. Kf2 Ne4+ 113. Ke2 Kf4 114. Kd1 Rg2 Here the computer stopped to play, hence obviously he resigned .In fact above there was not a 3-fold repetition, it was a 7-fold repetition of the same position. |
|
#15
|
|||
|
|||
|
On Apr 13, 2:47 am, Sanny wrote:
Now if you think harder you can win Beginner with a Pawn odds. Easy Level can win you with a Pawn Odd. meaning if you have a extra pawn then your game will be equal to that of Easy Level. Here is the new Rating I give to all Beginner: 1900 (5-10 sec/ move) Easy: 2100 (20-30 sec/ move) Normal: 2300 (1-2 min/ move) Master: 2500 (5-10 min/ move) It doesn't work like that. First of all, I am still winning most of the games; the draw problem nips me when I have gotten into lost positions, while in winning positions, I can and do win. I also still win a lot of drawable positions in the endgame. (Okay, I botched *one* ending in which I could have had a Q vs. B&N, with all pawns on one side of the board. Instead, I went for Q vs. R&N, thinking I had a forced win of a pawn with check. GetClub defended well and forced me to repeat the position.) Secondly, as we saw with the pawn-and-move odds match between Rybka and GM Roman Dzindzichashvili, when you start the game up a pawn the program can have severe problems in trying to draw, let alone win. In fact, most of RD's moves were aimed at trading pieces off, not improving his position or fighting for the initiative, etc. Add in some problems with the contempt-factor causing "issues", and it is very possible that the GetClub program could start off losing virtually every such odds game to me. Without the contempt-factor, I can just trade and win, thanks to my "free" pawn. Now, as for the 1900-2500 rating point range, I think you are getting much closer but still, these numbers are much too high. You will know when the program reaches 2500 by the fact that everybody will lose every game, except those who merely operate another chess program and enter the moves. Maybe the low end is a fair estimate-- 1900 USCF, but 2500 for Master level? In any case, you should begin seeing a lot more wins now, and as you already pointed out, more draws. -- help bot |
|
#16
|
|||
|
|||
|
When GetClub is at advantage, I have told it to change its move So
that repetition with draw do not happen. So After 2 repetition it changes its move to avoid draw. on move 86 (reptition number 1 is 87.Kf2, then 89.Kf2, then 91.Kf2, when making this move I should have had the option to claim a draw, then further reptitions: 93.Kf2, 95.Kf2, 97.Kf2, 99.Kf2, only then did the computer deviate): 8/8/8/4k3/8/4K1r1/7n/8 w In fact above there was not a 3-fold repetition, it was a 7-fold repetition of the same position. Yes, 7 fold repetition is not seen by it. And also It thinks Three moves repetition only when Computer has made three repition. Incase Human Opponents has made three fold repetition It do not make it a Draw. That needs to be corrected. If you are Strong Just win the game. Why are you loking for Draws? If you are weak just play with Beginner Level. As higher levels are much stronger. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html |
|
#17
|
|||
|
|||
|
initiative, etc. * Add in some problems with the
contempt-factor causing "issues", and it is very possible that the GetClub program could start off losing virtually every such odds game to me. *Without the contempt-factor, I can just trade and win, thanks to my "free" pawn. Could you tell me the weaknesses in the game. So that I may improve the GetClub Chess further. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html |
|
#18
|
|||
|
|||
|
On Apr 17, 1:35 am, Sanny wrote:
In fact above there was not a 3-fold repetition, it was a 7-fold repetition of the same position. Yes, 7 fold repetition is not seen by it. And also It thinks Three moves repetition only when Computer has made three repition. Incase Human Opponents has made three fold repetition It do not make it a Draw. That needs to be corrected. If you are Strong Just win the game. Why are you loking for Draws? If you are weak just play with Beginner Level. As higher levels are much stronger. Sanny, even a "strong" player can sometimes find himself in a bad position. Take Frank Marshall, for instance; a highly-rated player, this poor fellow is remembered today for his Q-g3!!! move, and for the fact that he repeatedly lost -- by clean scores -- to such players as Em, Lasker and Jose Capablanca. The point is not whether GetClub's opponents are weak or strong; it relates to the actual rules of play! The official rules by which we chess players must abide (except maybe for sneaky cheaters) state that a three-fold repetition of position, all things being the same (en passant capture and castling rights) and with the same side to move, allows a draw to be claimed. In fact, they annoyingly insist that such a claim must be made *before* making the move in question. These pesky rules and regulations are not necessarily endorsed or liked by everyone; but they are necessary in order to have a "game" like chess in the first place. -- help bot |
|
| Thread Tools | |
| Display Modes | |
|
|
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| We Are Your All GetClub All The Time Newsgroup!!! GetClub GetClub GetClub GetClub GetClub! | The tired of GetClub club | rec.games.chess.computer (Computer Chess) | 20 | February 28th 08 08:24 PM |
| rec.games.chess.misc FAQ [2/4] | pribut@yahoo.com | rec.games.chess.misc (Chess General) | 0 | May 23rd 06 06:24 AM |
| rec.games.chess.misc FAQ [2/4] | pribut@yahoo.com | rec.games.chess.misc (Chess General) | 0 | February 19th 06 06:44 AM |
| rec.games.chess.misc FAQ [2/4] | pribut@yahoo.com | rec.games.chess.misc (Chess General) | 0 | November 18th 05 06:36 AM |
| rec.games.chess.misc FAQ [2/4] | pribut@yahoo.com | rec.games.chess.misc (Chess General) | 0 | November 3rd 05 06:30 AM |