Hello everyone, even as a non-serious player, I can get a pretty easy win against the Computer-3-Hard chess computer on chess.com, (These computer series are on live chess). I did not end up winning the game, but I got a huge lead, and if I was a little better about not making blunders on chess.com, I could have won.
Note: I was playing the black pieces, so you should flip the board.
Here's the game, but be aware that the computer may not always do the same thing if you always make the same moves.
Here's how, watch the game: Computers Don't expect you to keep chasing their army around the board, in their searches, that logic isn't easy to program. That's how many chess players can sneak in perpetual checks, and amazing comeback wins against strong computers. That's the trick, this computer was rated 400 more than me, but I just calmly symmetrically moved, let the other army charge, and then chase the knight away until it forced white to lose a minor piece.
You probably won't encounter the same position, but you can try to use this game to help your (against computer play). One last time. Principle: Chase the enemy pieces around the board!
Note: I was playing the black pieces.
Here's a win in 12 moves vs the medium computer.
"Computers Don't expect you to keep chasing their army around the board, in their searches, that logic isn't easy to program. That's how many chess players can sneak in perpetual checks, and amazing comeback wins against strong computers. That's the trick"
This advice is rubbish: I'm no programmer but i'm sure it is easy to program, which is why we dont see Rybka playing such crap (that said, it just occured to me that by "strong" chess engine you might not have been refering to the likes of mighty rykba. If by "strong engine" you mean something under 1800 then you may have a point...But even then, this 'advice' would only ever be true if such an >=1800 had no access to a decent opening database.)
as for the game you post, both sides played a terrible game of chess: white completly fails to correctly use its central advantage, and black embarks on some rubbish wingplay. if you continue to play chess like that, then that computer will always be rated 400pts higher than you.
Wow Nice job totally destroying the computer orangehonda, cool checkmate too.
Computer Hard does some very odd things (http://www.chess.com/forum/view/livechess/computer-hard-sometimes-makes-bizarre-errors), however, Computer Impossible does not (as far as my meagre skills are able to expose). I think this is a problem with how the chess.com computer is dumbed down to patzer level. I've found that Fritz does an excellent job of dumbing down, without playing these nonsensical moves.
You can just leave this forum if you're going to criticize my advice. Also, why would you suddenly compare Rybka to my advice? Of course Rybka wouldn't make that move, it has extreme depths in its searches, and sometimes key positioning, by the way, I'm very knowledgeable on Rybka. Why would you assume that Rybka is programmed against my advice specifically just because it plays very good chess? And if you actually read the title, it said how to beat the computer on chess.com, not Rybka, and by the way, who knows if this technique was challenged to Rybka? I'm just giving the chess community some advice to play against the Computer-3-Hard.
If you want to really know something about Rybka, and all chess computers, here you go:
-Predicts exponential growth of possibilities
-Branches cut off if violates evaluation function
advantage (see below)
(There are many more variables
that effect the evaluation function,
such as isolated pawns, and some programs such as Rybka can even evaluate key positions, but the protection of the chase advice would be difficult due to the probability of interference with its own evaluation function, and in absence of the objective score for this rule.)
-If the score is negative, it means the
human is winning, if its positive, it means
the computer is winning.
You may not know this, but I've had stronger wins on stronger computers before, using this technique. Thank you for commenting though. I hoped you learned some new information :) By the way, I'm not arguing, just giving information, so you can leave this forum without feeling like arguing.
as for the game posted, both sides played wonderful chess, especially me, ha!
What the hell kind of opening was that ?!
//You can just leave this forum if you're going to criticize my advicehaha! "If you disagree with me and want to bring your reasoning to people's attention, please don't bother."//Computers Don't expect you to keep chasing their army around the board, in their searches, that logic isn't easy to program. What you're essentially saying is that computers aren't capable of putting moves where a piece is going to be forced to move again into their trees - which is possibly the most absurd thing I've heard stated with such conviction for quite some time.
You can just leave this forum if you're going to criticize my advice. If someone expresses a statement or idea on a chess forum, they should expect to be challenged. by demonstrating how bogus your ideas are i'm "helping" everyone who reads this thread.Also, why would you suddenly compare Rybka to my advice?you said "strong computer"...Rybka would fit my definition of "strong computer"by the way, I'm very knowledgeable on Rybka.the remainder of your post completly fails to demonstrate such knowledge. Why would you assume that Rybka is programmed against my advice specifically just because it plays very good chess?because chess engines always search for the best move and the best move in any given position is compeltly determined by the best reply (see: 'game theory')So, Rykba wont ever even move the knight in a position from which it be chased away in the first place (unless there is good reason to do so).And if you actually read the title, it said how to beat the computer on chess.com, not Rybkaconceded. I'm just giving the chess community some advice to play against the Computer-3-Hard. and i'm merely pointing out how "useful" your advice really is.want to beat hard computer? play solid chess. If you want to really know something about Rybka, and all chess computers, here you go:I'm affraid to say, you taught me nothing: this stuff just about everyone knows. "the protection of the chase advice would be difficult due to the probability of interference with its own evaluation function"Its own evaluation function will not have a "dont let the peices get chased round the board" rule, rather, it will be built into to its understanding of tempo, inititiative, space, peice position, etc. nobody programs a rule that says "always promote to a queen" because even the most basic of the computers evaluation function (Queen=8, rook=5, bishop=3 , etc) already understands that promoting to a queen is (rare exceptions aside) a good thing to do.I've had stronger wins on stronger computers before, using this technique. judging from the game you post, I genuinely doubt this. I hoped you learned some new information unfortunalty not. By the way, I'm not arguing, just giving information, so you can leave this forum without feeling like arguing.No. whether you intend to or not, you are in fact arguing.
Very interesting, nice job.
no, I'm not saying it's not possible, I'm just saying the chess.com computer does not use positional recognition (besides the opening), only possibility searches, sometimes its just a matter of a little logic, READ THE TITLE: how to beat the online chess computer.
you are continuing to try to find loop holes around grammar, nice try, you are no longer contributing to the thread, and I would please ask for you to no longer remain active on this thread, unless you will contribute ideas, instead of perpetuating my statement that 'you should leave this thread if you're just going to keep criticizing my advice [in an unrelevant rude mannar]' (the brackets added part of the quote that hopefully will make the statement more clear, and now you know, so one more time let me state it:
"If you're not going to contribute to this thread with relevance, free of rudeness, and misinterpreted criticisms based of the lack of comprehending the title, then please no longer stay active on this thread, thanks in advanced."
Also, I am not arguing,whether or not you interpret it that way. It is stubborn to interpret my text as arguing when I even made it clear that I did not mean for it to be that way, (there's a lot more to the text)
There is no need to program the computer to include "chasing their army around the board" in their searches. Computer algorithms create a list of candidate moves for each position, then examine all of the responses for both sides and evaluate the resulting positions. Using a min/max algorithm they find the best line for a given depth. The "chasing their army around the board" line is just one of millions of possible lines that the computer will eventually examine, given enough time.
it's not a computer principle, it's a strategical principle, however good attempt to transform that statement into computer language, but once again, it's not a series of numbers, it's a strategy. If you take a look at the evaluation function, and the in deph search 4 quadrent depths in Rybka, for example, the attack on pieces makes an immediate threat (unless there's pins) of that piece being captured so the computer concentrates more possibilities on that, by attack the same piece multiple times, it causes an extreme depth search that only overlaps the other searches of the same general algorithm, this is not true for strong computers such as Rybka, but it is true for perhaps Computer_3_Hard, with only minor search engines, ONCE AGAIN, I have to say, read the title
it's an attempt at single pawn symmetry, please don't use that word
I am very knowledgeable on Rybka, and I do succeed to express such knowledge, but that's not the purpose of the post, it's for chess.com computers, and I'm starting to doubt the counters to my advice, because apparently the counter-advice people haven't even been comprehending the title! It's silly
"Computers Don't expect you to keep chasing their army around the board"
This advice is rubbish: we dont see Rybka playing such crap
Blackadder: Please provide us with an illustrated game (or two) showing how Rybka does not fail for the same anti-computer trap.
you did learn a lot of new information, you just won't admit it