Can Humans Evolve to Beat the Engines

[heroku]

How long in human years would it take to evolve into beating the engines and comprehend and understand chess in a vast scale. In ever growing chess content and education, it should be doable within 7 years I presume.

[towforce]

How long in human years would it take to evolve into beating the engines and comprehend and understand chess in a vast scale. In ever growing chess content and education, it should be doable within 7 years I presume.

The bad news: humans will never be able to beat chess computers using current chess knowledge.

The good news: uncovering new chess knowledge (deeper patterns in chess) would enable humans to easily beat today's chess computers.

More bad news: the chess computers would acquire this new knowledge and get back on top again.

[jefk]

uncovering new chess knowledge (deeper patterns in chess) would enable humans to easily beat today's chess computers.

well if that's your opinion, i don't think that's correct.
Imo it's certainly would not be possible to 'easily beat (a comp)' (if at all possible); just like
with calculation eg. multiplication of numbers, the machines are vastly superior nowadays.

PS yep i know about ideas with much better pattern recognition (eg neural nets?)
you would need less calculation, but also for pattern recognition there are human
limits; and usually quite different, eg. between Carlsen or a beginner.
For the comps there also are nnue (or other technical sorts of) limits, but anyway
uncovering 'new' (more) chess knowledge would probably be done earlier again
by some comps than by humans.
Third we already almost have reached 'draw death' (certainly in high level
correspondence chess) so basically there's fundamentally not much to improve
anymore in normal games, at least not from normal openings.
example:
https://www.iccf.com/event?id=89454

some could think (or argue) that this is because the top engines like SF 17 now are
all equal strength and that in theory a much stronger engine could still win more games.
Not so, if you have enough experience with such correspondence games (and opening analysis)
you would know we are reaching a fundamental (draw) limit, where a top engine cannot lose
anymore, at least not with sufficient time and hardware (which doesn't have to be so much
nowadays btw). Sure some engines can still be improved, but they will not win more
games under such conditions (for humans it will be quite some time before someone
would reach such a limit, look at the Elo of the human vs top comps).

https://www.facebook.com/groups/iccf.fb/

[JohnW]

How long in human years would it take to evolve into beating the engines and comprehend and understand chess in a vast scale. In ever growing chess content and education, it should be doable within 7 years I presume.

Not a chance, Stockfish is already +800 points above Magnus Carlsen and the gap is only getting wider

[heroku]

Only 800 points? I thought at least 2000 points ahead engines from humans. I am also talking about natural evolution to be able to organically expand your intelligence and comprehension skills. And the #1 fault in human chess is they make mistakes if Super GM's never make mistakes due to emotional tension, using their synthesis of all the chess literature and applying them strictly without mentally cracking into mistakes, they can at least draw engine's if not win by being able to come up with some sort of conceptualization of chess just like how they can invent the engines, do ten different things at the same time, while engines hack their clogs.

[towforce]

...Imo it's certainly would not be possible to 'easily beat (a comp)' (if at all possible); just like with calculation eg. multiplication of numbers, the machines are vastly superior nowadays...

It might be that today's top engines are close to being unbeatable - but they may have a weakness that makes them beatable. Suppose:

1. deep chess patterns which are easy enough for humans to learn were uncovered which reliably told the outcome of a position

2. more deep patterns were uncovered which made it easy to recognise which moves would lead to complex positions

Playing chess would then become:

1. Make a mental list of moves that improve, or at least don't worsen, the win/draw/lose situation

2. From these moves, choose the one which leads to the most complexity

A complex position:

1. Weakens an engine's ability to accurately evaluate the position

2. Throttles the depth of the search tree, preventing the engine from looking a long way ahead

Then today's engines might be beatable by humans.

Of course, it wouldn't be long before the engines received these deep patterns: the advantage would be short lived.

[heroku]

...Imo it's certainly would not be possible to 'easily beat (a comp)' (if at all possible); just like with calculation eg. multiplication of numbers, the machines are vastly superior nowadays...

2. more deep patterns were uncovered which made it easy to recognise which moves would lead to complex positions

2. From these moves, choose the one which leads to the most complexity

A complex position:

1. Weakens an engine's ability to accurately evaluate the position

2. Throttles the depth of the search tree, preventing the engine from looking a long way ahead

Then today's engines might be beatable by humans.

I think human's should go in reverse, there is no way human's can comprehend complex positions better than the engine, for example which Super GM can solve a mate in 15, it's just impossible and also unnecessary. So human's should seek to recognize the positions that lead to conceptual simplicity without mind boggling complications, but somehow make the engine's confused, it might lead to at least more draws.

[Uri]

I think that chess engines like Stockfish and Komodo (for example) still suffer from the same kind of weakness: there are many instances in which the engine gets a position that he doesn't 'understand' well and this triggers a search explosion.

[syzygy]

How long in human years would it take to evolve into beating the engines and comprehend and understand chess in a vast scale. In ever growing chess content and education, it should be doable within 7 years I presume.

It will take longer than 7 years for humans to start growing graphics cards in their brain.

[heroku]

How long in human years would it take to evolve into beating the engines and comprehend and understand chess in a vast scale. In ever growing chess content and education, it should be doable within 7 years I presume.

It will take longer than 7 years for humans to start growing graphics cards in their brain.

Are you serious, human vision is in 60GigaK definition.