10 Oct 07
Originally posted by MrJohnnot necessarily...
So, would this imply that a grandmaster level correspondence chess player would have no problem whatsoever with Deep Blue?
After all, we have plenty of time to check for blunders.
Thx! :-)
McDonald and Spassky are both GMs, but i'd have to say that, if they were to play when spassky was in his prime, Spassky would win every game.
It depends on the strength of the GM
we'd need a Super-GM
let's analyze:
computers are basically programmed GMs plus laser tactical sight...
tactics come from mistakes on behalf of the opponent.
so, if the GM is given enough time to look for and correct such mistakes, then they would draw.
this may explain the huge number of draws in today's GM circuit. chess is becoming such a precise science, these mistakes are few and far between.
10 Oct 07
Originally posted by Pigface1There's a removed post in the beginning, before the 3rd in this thread. Doesn't say why.
Thats odd, I had a post in this thread, could whoever removed it (I assume it was removed) please let me know why? I am pretty certain I didnt break any rules and only want to know why so I wont inadvertantly do it again.
10 Oct 07
1 edit
Originally posted by MrJohnabsolutely. deep blue was quite a long time ago.
So, would this imply that a grandmaster level correspondence chess player would have no problem whatsoever with Deep Blue?
here's what kasparov said about the results of the pal/css freestyle tournament of 2005:
"Even the strongest computers were eliminated by IMs and GMs using relatively weak machines to avoid blunders. Three of the four semifinalists were grandmasters working with friends and computers."
which was also exactly what he had predicted to happen. lone machines had no chance.
10 Oct 07
Originally posted by wormwoodBut what I found funny was that:
absolutely. deep blue was quite a long time ago.
here's what kasparov said about the results of the pal/css freestyle tournament of 2005:
"Even the strongest computers were eliminated by IMs and GMs using relatively weak machines to avoid blunders. Three of the four semifinalists were grandmasters working with friends and computers."
which was also exactly what he had predicted to happen. lone machines had no chance.
"The computer-assisted PAL/CSS Freestyle Chess Tournament, staged on Playchess.com, ended with a shock win by two amateurs: Steven Cramton, 1685 USCF and Zackary Stephen, 1398 USCF, using three computers for analysis, defeated teams of strong grandmasters all the way to victory in the finals. We bring you a first flash report with games and results."
If computers play such bad chess, why was tournament won by two amateurs with comps, rather than GM's and IM's?
10 Oct 07
Originally posted by MrJohnThat brings up an interesting point. If a GM played a top-ranking engine in a correspondence game, with, say, a week per move, I personally would bet on the engine. After all, if you can let the engine analyze the position for an entire week, it would be able to cover an unfathomable number of lines and variations, and their respective tactical and positional qualities.
So, would this imply that a grandmaster level correspondence chess player would have no problem whatsoever with Deep Blue?
After all, we have plenty of time to check for blunders.
Thx! :-)
10 Oct 07
Originally posted by toygeryea, what an incredible achievement that one. 😀 my guess is that freestyle chess has some differences to other types of chess, such that specializing in it will yield some special advantages that neither an engine nor a unprepared GM can necessarily provide? what I read about the zacks team was that they had played countless freestyle games, and studied extensively on databases. making them essentially specialists in the variant (and remember, it was the first tournament, the GMs were probably not very well prepared for freestyle) and that along with 'perfect play' might have lead them to victory straight from an opening advantage. to me it seems plausible that a master would think in a position: "okay, that gives my opponent a slight advantage, but not enough to win" and be proved wrong with the more accurate centaur play than he would ever have been used to.
But what I found funny was that:
"The computer-assisted PAL/CSS Freestyle Chess Tournament, staged on Playchess.com, ended with a shock win by two amateurs: Steven Cramton, 1685 USCF and Zackary Stephen, 1398 USCF, using three computers for analysis, defeated teams of strong grandmasters all the way to victory in the finals. We bring you a first flash rep ...[text shortened]... uch bad chess, why was tournament won by two amateurs with comps, rather than GM's and IM's?
but it all still doesn't change the fact that every single engine without human assistance got beaten. even the strategical understanding of a 1300 & 1600 player was enough for that.
looking at the latest pal/css freestyle tournament, the best machine was 'mission control' with its 32 processors of brute force. it came in sixth with a -1 score. the winner was rybka + its IM author and his IM wife, ending up with +3. in 10 games, that's quite a lead.
10 Oct 07
Originally posted by wormwoodI very much doubt there are many positions where a 1300/1600 player will improve upon Rybka either tactically or strategically. Their skill was the coordination of their 3 analysis machines; getting the PC to step deeper into given lines; preparation of the opening book (as you mentioned); choice of opening with regards opponent, etc.; time management for the analysis; etc.
even the strategical understanding of a 1300 & 1600 player was enough for that.
Do you think they were having discussions such as "the computer wants to advance the pawn but it leaves too much of a weakness"? What's the chances of them being wrong versus them being right?
In short, they won because they leveraged the 3 PCs and software very effectively.
10 Oct 07
Originally posted by Varenkawell, naturally I don't know what they actually did. but they played as a centaur, not computer only. what else could they do but make strategical choices? certainly not tactical, and the computer can handle the opening with their book.
I very much doubt there are many positions where a 1300/1600 player will improve upon Rybka either tactically or strategically. Their skill was the coordination of their 3 analysis machines; getting the PC to step deeper into given lines; preparation of the opening book (as you mentioned); choice of opening with regards opponent, etc.; time management for t ...[text shortened]... right?
In short, they won because they leveraged the 3 PCs and software very effectively.
10 Oct 07
Originally posted by wormwoode.g. on one machine they see that the engine has two close top options. They expand each on these on the other two machines, while the first machine continues to verify that no other candidates rise to the top. Or they use a different engine to get it's opinion of the candidates. etc. e.g. trying to get the engines to refute their own suggestions
what else could they do but make strategical choices?
It's this sort of work that I'd expect they would be doing, and not looking for the very rare occassions where they expect to know better than Rybka.