Why do GrandMasters Lose almost every Game vs Top Best ChessEngines ?

Sort:
EloInfinity

I have found many reasons on the Internet but of all what is the truth. ??

Since a GM Knows everything about Chess and do it very well , where do they go wrong in the games vs Top Chessengines ?

u0110001101101000

If you could win a game of chess by knowing more, we wouldn't have 20 year old world champions, and players wouldn't peak in their 30s.

Knowledge is important, but tournament chess is about performance. Programmers have included enough in the eval functions of computers that their better hardware means they out preform humans in games.

They're still not completely better in analysis, where humans can still find better moves based on general knowledge, but in a game the human will make too many errors to win.

u0110001101101000

I like to think of it like writing a novel. A human can tell a great story... but the first draft is bound to have typographical errors, grammar and punctuation issues, spelling errors, etc.

Then imagine a computer program that writes. It will tell a s**t story, but the grammar and punctuation will be perfect.

In chess, the little things count a lot, so the engine wins... but it still plays like an idiot

EloInfinity

If i Understand corectly you mean Human GM's Lose to Top Chess Engines for reasons that are not based on Knowledge ,

U mean they lose because of :

Mental Fatigue , Mind Concentration , Awakeness ecc ecc ??

Megabyte
0110001101101000 wrote:

In chess, the little things count a lot, so the engine wins... but it still plays like an idiot

To be honest, the latest engines can play in a really simple and clear way. Humans usually want to overcomplicate the game when it can be easily won by improving space, or even doing some simple trading. Computers aren't bound by fancy competitiveness, so they win by simple, strong objective and brute-force analysis.

Reb

Engines dont " know " anything about chess just like calculators don't " know " math but are still good at it .  Cool

u0110001101101000
Megabyte wrote:
0110001101101000 wrote:

In chess, the little things count a lot, so the engine wins... but it still plays like an idiot

To be honest, the latest engines can play in a really simple and clear way. Humans usually want to overcomplicate the game when it can be easily won by improving space, or even doing some simple trading. Computers aren't bound by fancy competitiveness, so they win by simple, strong objective and brute-force analysis.

At least in winning positions, computers are more likely to over complicate things, and humans are very good at simplifying. In fact this is one easy way to tell when you're playing an engine (they don't play simply).

Fancy competitiveness? Sounds like something you made up. Humans error because we use what amounts to generalizations. So when engines play something bad, that "miraculously" turns out to be ok, it's because we judge based on knowledge while they calculate lots of long lines.

When they run into the horizon effect, then their silly moves turn out to be actually silly. But yes, like you said modern engines are much better and this is less common.

u0110001101101000
EloInfinity wrote:

If i Understand corectly you mean Human GM's Lose to Top Chess Engines for reasons that are not based on Knowledge ,

U mean they lose because of :

Mental Fatigue , Mind Concentration , Awakeness ecc ecc ??

 More or less.

Basically engines calculate a lot, while humans use generalizations and yes, also fatigue.

Megabyte
0110001101101000 wrote:
Megabyte wrote:
0110001101101000 wrote:

In chess, the little things count a lot, so the engine wins... but it still plays like an idiot

To be honest, the latest engines can play in a really simple and clear way. Humans usually want to overcomplicate the game when it can be easily won by improving space, or even doing some simple trading. Computers aren't bound by fancy competitiveness, so they win by simple, strong objective and brute-force analysis.

At least in winning positions, computers are more likely to over complicate things, and humans are very good at simplifying. In fact this is one easy way to tell when you're playing an engine (they don't play simply).

Well, it's not that they overcomplicate things. They always choose the quickest win, even if the quickest win means it would be harder for a human to play (e.g, giving out a queen for seemingly no reason when giving it up will mean mate in 10, instead of mate in 20). But more often than not in the opening and middlegame, computers will pick solutions that seem surprisingly simpler – Stockfish and Rybka are known for that.

I don't really have a word for human "competitiveness". Basically, I mean that a human will play illogically because they don't like a certain move, or because they find a certain mate pattern pretty (regardless of actual effectiveness), or because they want to make their opponent suffer, and so on. And these tactics will obviously backfire.

I believe one example of that would be men playing against women in chess. According to what I have read on chessbase (http://en.chessbase.com/post/male-che-players-show-elevated-aggreivene-against-women), if both of them have exactly the same rating, men will play more aggressively than women, and this sort of illogical play will make men take unnecessary risks, and play slightly worse (1% worse). As I said before, computers don't go through these psychological issues at all.

Megabyte

Oh, and this too: https://www.chess.com/forum/view/general/castling-is-for-sissies?lc=1#last_comment. It's a thread that states "castling is for sissies".

Now, of course this is a joke thread, but I see a human thinking within a match that castling would be a "cowardly" move and dismiss it wanting to attack the opponent and losing when castling would be the best option. It's things like this humans are subject to.

u0110001101101000

Oh, I think I see what you mean by competitiveness now.

As for choosing simple ways of playing in the opening, the only times I've encountered that myself are when the engine is unwittingly choosing a way of playing that removes winning chances. A human may play that way for a draw, but it's not a practical way for humans to play against other humans.

I don't think of this as psychology, more like being practical (and engines having no concept of winning chances or how much play exists in a position).

Alchessblitz

Imo a GM never knows everything about chess and chess is a difficult game and not masterable for a human.

In short a human (even if very strong) is potentially capable of making lots of mistakes or bad moves and strong bots are formidable punishers of errors and bad moves.

tygxc

Tactics. Engines err less and punish human error.

Alchessblitz

Indian Derby: Gukesh v Vidit! Table Topper Anish Faces Ding With Shared Lead | Tata Steel 2024 Rd 9 (youtube.com) On this video there is the game Parham vs Warmerdam which is very telling to understand "chess by humans" because when we ask the question of the topic I think it is because we have lots of clichés or false ideas of the game of chess played by humans.

In the game "at the exit of the opening" Parham is in a losing position which means that against a strong bot he can already give up but against a human it is not as obvious because for a human there are more parameters that come into account such as the parameters of psychology, fatigue but also that the game of chess is difficult and a human can potentially make lots of mistakes.

CRAZY_TEEN

idk

VenemousViper
EloInfinity wrote:

I have found many reasons on the Internet but of all what is the truth. ??

Since a GM Knows everything about Chess and do it very well , where do they go wrong in the games vs Top Chessengines ?

"Since a GM Knows everything about Chess and do it very well" GMs make mistakes just like humans. Do not be fooled by the fact they often make brilliant sacrifice and seem to always be right.

"where do they go wrong in the games vs Top Chessengines ?" Well, you see, engines are better than humans because of consistancy. If there is a forced mate in 19 moves, an engine will find it. If there is a winning tactic, an engine will also find it. Every time.

chessterd5

Engines have the programming and information storage ability to calculate multiple lines to the depth that they are programed for in a time frame that is impossible for humans to do as well. But if you give a human the time neccesary and information resources that a computer has to calculate he can be competitive.