Author: José de Jesús García Ruvalcaba
Date: 12:15:06 07/24/99
Go up one level in this thread
On July 24, 1999 at 02:14:38, Melvin S. Schwartz wrote: > >On July 24, 1999 at 01:36:56, Mark Young wrote: > >>On July 24, 1999 at 01:20:55, Melvin S. Schwartz wrote: >> >>> >>>On July 24, 1999 at 00:52:47, Dave Gomboc wrote: >>> >>>>On July 24, 1999 at 00:14:32, Melvin S. Schwartz wrote: >>>> >>>>>The game as I stated in my original post was at 40/2. Hiarcs selected a4 and >>>>>after making the move its score dropped to a minus. I therefore allowed Hiarcs >>>>>to replay the move. Hiarcs then selected Rf1. I again let Hiarcs replay the move >>>>>and again it selected Rf1. >>>>> >>>>>Now, I believe it is quite evident that Hiarcs learning function did not allow >>>>>it to replay a4 and insisted on playing Rf1. If you think a4 was not a bad move, >>>>>then how come Hiarcs would not play it again and insisted on Rf1? >>>>> >>>>>By the way, your headline in the post to me is not appreciated. I expected an >>>>>intelligent discussion about this and not some silly rhetoric. You asked for >>>>>examples and I gave you some. You insist on dwelling on this trying to dispute >>>>>even the program you are raving about. Hiarcs recognized a4 was bad and that's >>>>>why it wouldn't play it again - unless you have some better explanation? >>>>> >>>>>Regards, >>>>>Mel >>>> >>>>Did you play a move in reply after it played a4 before returning to the position >>>>before a4? If so, which? It could be that Mark's copy is ____________ >>> >>>Hello Dave, >>> >>>I did not respond to a4 because I saw the score drop from a plus to a minus and >>>decided to let Hiarcs replay the move - I was curious because a4 appeared to be >>>a bad move as the score indicated. When replaying the move, Hiarcs selected Rf1 >>>with a plus score. I then let it replay the move once again and again it >>>selected Rf1 again with a plus score. I decided to let the game continue with >>>Rf1. Apparently the learning function prohibited it from playing a4 again >>>because it was a bad move. Otherwise, what could possibly be the explanation? >> >>I can tell you this it had nothing to do with hiarcs 7.32 learning fuction. You >>played the game at 40/2. I let it think for over 20 mins on this position and it >>never showed at - score with the move a4. >> >>And if it did, it does not matter, you claim the move a4 was a blunder. Why >>because Hiarcs 7.32 plays the move a4 but on your system it shows a - score. >______________________ > >This is it for tonight because I have to get to sleep. I stated many times that >it was after Hiarcs selected a4 and was waiting for my move its score dropped >from a plus to a minus. It was quite a drop and so I let it replay the move and >with Rf1 it regained a plus score. Since when is going to a minus a bad move you >ask? I would say that if you have a slight advantage in the score and you go to >a minus, then you played a bad move. There are other posible explanations: 1. The program was already in a poor position and had not realised it before. 2. The score can have big swings when going from one ply to the next, specially in programs that do not use null-move (I do not know if Hiarcs is a null mover). 3. The evaluations are wrong. 4. Some root porcessors may never realize that a move is bad before actually playing it (I do not know if Hiarcs is a root processor). >It may not mean the game is lost, but if >another move such as Rf1 keeps the plus score - then a4 was a bad move. That is >my opinion. If you disagree, fine! I am not going to drag this out. Mark, I have >too many things to do to keep this going. It's after 2 am and I should have gone >to bed hours ago. My wife has been waiting and... Have a good night. Maybe we >can continue tomorrow, though it appears we are not going to convince the other >to cause an agreement on this issue. > >_____________________________ > >>Since when does a - score mean a move is a blunder, just because a program >>thinks it is. I Looked at the position as a human chess player, and analyzed it >>myself. I could not find a way to refute the move, nor could the other chess >>programs that played against Hiarcs 7.32 from this position. >> >>> >>>Regards, >>>Mel >>> >>>not analyzing a >>>>particular response as deeply as yours did, not having been prompted to do so by >>>>a move from the opponent. >>>> >>>>Dave
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.