Author: Gerd Isenberg
Date: 09:24:56 01/25/02
Go up one level in this thread
On January 25, 2002 at 06:51:36, Ulrich Tuerke wrote: >On January 25, 2002 at 05:57:42, Gerd Isenberg wrote: > >>On January 25, 2002 at 02:36:48, Ignacio Urbelz wrote: >> >>>I could not connect to ICC to see the games live where there might have been >>>some explanation to this, so I post the question here. Anyone can help explain >>>the disater results H8x is getting in the tourney ? Some of the lost games have >>>really strange moves. Is this a bug ? Thnx >>> >>>Ignacio Urbelz >> >>In the first round of CCT4 IsiChess had black againts Hiarcs8X. I operated >>IsiChess manuelly by myself. >> >>In this position, short after Bookline was finished and both programs had to >>calculate, Hiarcs8X played 22.Rxd7 instead of 22.Nd4. >> >>[D] 2kr1b2/3b1p2/p3pPr1/1Nn1q3/7Q/2N5/PPP4P/2KR1B1R w - - >> >>I was very confused an nervous - Isis score immediatly went to +2..3. Is there a >>trap? What's going on here? But a few moves later Hiarcs8X operator saw that >>something went wrong on his Board - there was stil a Black Bishop on d7 on his >>Board. Obviously there was an interface problem. A wrong 22.move was posted by >>Hiracs8x. >> >>After asking folgers and deepFritz, what to do, i offered to replay from move >>22. But there were already discussions in the Channel Forum - so Hiarcs8x >>Operator decited to resign due to possible later protests. >> >>I felt shamed about the circumstances, IsiChess won the game, specially not >>asking my opponent directly after 22.Rxd7, "do you really want to play this >>move?". >> >>Gerd > >Well, after all it's supposed to be chess, what the computers are playing, and >there is a fixed set of rules for human chess; the same applies to computer >chess. Withdrawals of moves are not allowed. > >I admit that it's a pity when it's the interface which produced the wrong move. >I remember a game played on a wmcc by ChessTal, where the engine had produced >the winning move (promotion to queen) but the GUI had chosen a knight instead. >Same thing. I think there is no other chance than considering the engine and its >complete sw/hw environment which is necessary to produce the move on ther server >as one unit. > >Otherwise, we could think of a lot of cases, where decisions will become >extremely difficult. > >Anyway, no doubt that you are a real sportsman, Gerd ! > >Uli Thanks Uli, I remember Jakarta 96, where two german programs were involved, i think Schach 3.0 and Heureka. So roundabout move 60 Schach 3.0 Operator (a nice women, but no chess neither computerchess expert) entered a wrong move. A few moves later they recognized that something was wrong. The difficulties were, that first the user interface from Schach was a real rudimentary character oriented one, and second the chessboards had no coordinates and third, after it was happened, Heureca was not able to take back moves or even to load a saved pgn-game due to a bug with ep-moves. Mr. Fisher(?), the Heureka operator, had to reenter the whole 60 moves. The situation was handled by Jaap quite professionally and after one hour or so the game could continue. So this was a human interface error - and no online tournament, of course. I think, as a consequence of this, the ICCA made some restrictions for operators in further ICCA-events. In the 95 match between ChessSytemTal (operated by Thorsten) and XXXX, i have my doubts until today, wether the decision was right. The grafical board from ChessTal showed a knight on the promotion square (only a wrong piece was drawn), but in several text windows the correct queen promotion was shown (game notation and PV with mate in 2 or so, instead of -mate in 3). Maybe arbiters decision was influenced by Chris Wittingtons ICCA-reputation. Even if both programs dont't play for the title, such "wrong" result may influence standings of others of course. In my opinion there is a difference between such obvious "interface" errors and real bugs by the enginge, producing bad moves. OK, difficult to decide. For me it would be fair, to take back one move, eg to immediately correct a wrong entered move because your mouse is zicky. (Is there a clear ICCA-rule about entering wrong moves?) If this is right, it should be common to inform your opponent, if his last move looks like a blunder to you or your engine, at least if you play manually. I didn't - and i played manually. With some potential to cheat a bit, eg. forcing easy moves, or waiting to enter opponents move, if your engine ponders this move, and you feel your engine may need some more time in a critcal situation. That's not (so easy) possible in real tournaments, but when you are alone at home, may be frustated by the game before and now a possible win, if your engines has enough time to find the right move... After my experience with the first CCT4-weekend, I strongly agree with those, who claim that playing an online tournament requires autoplay and a kind of common logfile. To acknowledge the last posted move by the protocol would be a nice service. Gerd
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.