Author: Stephen A. Boak
Date: 20:46:56 06/27/01
Go up one level in this thread
On June 27, 2001 at 21:48:44, Robert Hyatt wrote: <snip> >Sure... but I don't believe it is legitimate to say that a computer is 2900 at >tactics based on that... A 2900 tactical player would be very difficult to >handle. Some comments here, nothing necessarily very new in concept-- Let's say Kasparov, when he is in his most brilliant form, is a 2900 player. Maybe in reality it is 3000, but who knows, after all he defines the top of the playing scale, in general. In many games he plays against top opponents, he doesn't succeed in obtaining a position where his brilliance can be shown. However, he tries to steer many or most games (when he seeks a win to soothe his troubled ego) into situations where he can outwit his opponents and that brilliance can flash in some manner or other. What I mean is he seeks to steer the game into positions where he understands them better than his opponent, believes the possibility of winning exists, where his position is better, from which he can find & spring a tactical wonder on them when they are least expecting. [I guess it is actually brilliant, if you can see the tactics and a top class opponent cannot.] The point is he cannot always achieve those delightful positions, where a stroke of tactical brilliance decides the game--even if he is in 2900 tactical playing form and carrying a fighting mindset that slobbers to find a clinching shot! A comp may be 2900 tactical strength, but it doesn't know how to (consciously, inentionally) steer the position into ones where its own tactical brilliance may flare up and win the game. It simply steers itself where the search & eval think the score is highest, not necessarily into the richly complex positions that is brain candy to a Kasparov-type of player who wishes to outwit his opponent. It may outplay its opponent, but can't easily seek out the richest veins where mining for gold may contain a brilliant strike. I'm not saying it couldn't be programmed to achieve this more often. But I suspect comps aren't programmed this way now. There may be exceptions. At least, the comp may be less likely to be able to achieve such rich positions, against scheming, anti-comp minded GMs. :) If the GMs can keep the position's complexity from getting out of control, they will not be surprised by the 2900 comp's tactics--because there are no such brilliant strokes (hidden from the GM perception) in the positions the comp obtains. Kramnik trading queens early with Kasparov may be a great example of this in the human vs brilliant human arena. I agree, a 2900 tactical strength opponent would be very hard to handle, let alone beat, but if the positions are kept relatively 'sterile', then the opportunity for near-genius or genius-level tactics will be minimized. What seems to me to be tough about a 2900 tactical player (human or comp) is that you could make the slightest positional error, barely or not even perceptible to slightly lesser geniuses (top players), and the brilliant stroke suddenly lays bare not only the weakness (perhaps already noted by the owning player, perhaps not) and shows, proves, not only that it truly existed, but moreover how it may be fruitfully dissected, and that the weakness wasn't 'manageable' after all (in cases where the weakness was knowingly accepted in the tradeoffs inherent in fighting for a win--and taking commensurate risks). I think a weakness which is noted by a GM in his own position, is merely a positional weakness, a concept not realized concretely, until another, more brilliant GM dissects it with a tactical proof. Then we see that it was actually a tactical mistake, not just a positional weakness that was noted but thought to be manageable. These are just semantics to describe my wandering thoughts--not trying to create a whole new vocabulary. Tough to handle means to me that you could not accept any 'normal' slight weakness, in pursuit of a win or initiative or counterplay, without the fear that the 2900 tactical player (human or comp) might immediately, without warning, show you the truth in the position--that the weakness isn't so slight after all. --Steve
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.