Author: Mogens Larsen
Date: 15:19:09 06/12/00
Go up one level in this thread
On June 12, 2000 at 17:38:16, Dan Newman wrote: >What you're talking about is something that I think any of us chess programmers >would like to do, but is a problem that's a lot more complex to solve than it >may seem to you. What you seem to be asking for is a program that could >(effectively) generate opening theory (nearly as good as or perhaps exceeding >current opening theory), on-the-fly, during a game. This is asking for a >program to do a lot more than a human being can do (even Kasparov) in a realm >in which humans are (inherently) much better than computers... Opening theory >has evolved over decades, even centuries, with thousands of human minds working >on it, analyzing in a fashion that we've been unable to reproduce in a computer >as of yet. I'm aware that the problem is very complex, even though I know very little about programming. There's little chance of producing such a program in the very near future, but I was mostly adressing the problem with ignoring important parts of a chess game. There's no need to generate complete opening theory in realtime at this point. Maybe a step forward would be an advanced learning capability as an intermediate solution? Improving by playing. >I think that it's likely that such a program (if we are smart enough to be >able to write one) could also be given similar abilities in the middle and >endgames and would likely be completely un-beatable by a human. That would probably true, even though the middle game would be a tough enterprise. >Besides, if there were a rule created that computers can't make reference to >an opening book file on disk, programmers can simply put their opening books >into the source code as a table of data. There are many tables of data that >chess programs make reference to and this would just be another one (albeit >a large one)... I'm not against opening books in general, but the current possibility of mass storage could make it ludicrous. >The way I see it, when a chess program consults its opening book it's sort of >analogous to a human relying on his long term memory of opening lines that >he's studied (admittedly not a perfect analogy--but then computer chess players >aren't pure analogs of human chess players in any other way either). That's very true. Thank you for an exhaustive response. Best wishes... Mogens
This page took 0.01 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.