Will humans ever be able to beat computers at chess?

This will happen the same moment google translate starts working ideally 100% of the time

i dont think so. chess engines are improving faster than humans and chess engines are already superior to humans. when players become cyborgs to improve their play everthing will change

I can hit the computer with a hammer until it stops working

typical nigger

There are more possible chess games than there are synapses in your tiny little brain. Chess might be finite, but it’s not computable yet

did you feel smart writing this?

The more time you give a computer, the more cemented it's advantage becomes.

I think the dominance of computers will lead to an overall reduction of interest in chess, and other board games. It's not interesting to play a game with such deep theory. Consequently, the strength of chess players will gradually decline, and computers will leave us further behind.

These days I only play Fischer-random and Shogi, and only against human opponents. Chess is full of autists who can win simply by learning more theory.

10 person teams vs 1 computer?

Except chess is more popular than ever before

It really doesn't matter how many people you include.

There's also more people than ever before, and the number of people with internet access is increasing rapidly, what's your point?

Every high-level game is recorded, and every new novelty is added to theory. Chess is becoming a memory game, and memory games are fucking shit.

>a memory game
At super high levels, sure. So 0.001% of population is affected by that. The rest of us play intuitively because that's the only good strategy we have.

so 99.999% is getting its ass whooped 100% of the time by that other 0.001%. Cool game bro

That ship has sailed. Unless we augment humans with computers.

>At super high levels, sure
No, just at any ordinary meet of people rated higher than 1700, of which there are millions.

I played chess competitively from ages 9-17, then began losing all the time to people with less experience who just knew more theory.

That's why I only play Fischer random and Shogi, even at competition levels, the theory is extremely shallow.

>the strength of chess players will gradually decline,
I don't agree with that at all. If at all, 2 players play the best possible moves, the game is a draw. That's the future of human vs machine chess. Humans will get stronger because they have the advantage of computers in analysis.

>If at all, 2 players play the best possible moves, the game is a draw.
You have absolutely no way of knowing that, computers haven't completely solved chess yet.

chess.com/forum/view/general/true-or-false-chess-is-a-draw-with-best-play-from-both-sides

It's a draw m8.

No. And now neural networks are crushing the old chess engines.

I don't understand why people diminish the game itself just because computers are better or have solved it. Checkers/draughts - a solved game still has an active competitive scene in spite of it.

Chess players at the highest levels in general do get better from computers through more thorough and extensive opening preparation but that's far from a full game and encapsulates a already huge number of lines for just the opening

>I've been thinking about this.
you call that thinking?

>Chess is full of autists who can win simply by learning more theory.
This is not true at all. This tells me that you're stuck around 1800 elo and you have no idea how to play, but you're mad about people who beat you.

yes, but only when chess is solved

Leela zero plays like an über human rather than an engine

I assume that you mean 1800 ELO on some online system (meaningless)
1800 ELO FIDE or UFC is quite strong

Meant USCF not UFC

Yes.
Wrong.
The singularity is coming.

>1800 ELO on some online system
No.
>USFC elo
That's not elo, that's rating. Dumb american, as usual.
>1800 elo is quite strong
And this is how I confirm you're shit.

My main problem with chess is that when i play by intuition I do all kinds of blunders and when I try to calculate all the possible moves to avoid blunders then I feel like a real shitty computer which can analise like 5-10 positions per minute while even my phone can do several hundred.

Ironically engines and AI have made chess MORE popular.
A lot of CS people play chess and then work on Stockfish or try some kind of machine learning to make their own engine.
Not to mention the cool data analysis you can do with gane data.

You shouldn't calculate all possible moves to avoid blunders. That's what you're doing wrong.

>I've been thinking about this. Take a game without randomness: since it has a finite number of states, we can calculate the most optimal routes to win

This is understandable but actually false. Its true that chess has a finite number of states, but the universe is also finite. Turns out there are more possible chess states then can be explored inside of the finite universe. Many games can be fully explored but many others cannot, including chess.

That shit was trained by an engine and uses terabytes of data where its stores its neurons and weights and connections. Basically the only interesting stuff about it is that it needs to do less calculations to arrive at the same conclusion as an engine but only because all of those positions were pre calculated when it was trained by mega expensive tensor hardware and stuff.

When you see AI beating an engine like Stockfish it happens because SF runs on limited hardware and it just doesn’t get enough time to find the best move. Like when AlphaZero beat SF it happened because SF run on like a shitty laptop.

?
Elo is a rating system and USCF has used it since the 60s.
In fact USCF has used Elo since even before FIDE did

1500 is about the average Elo rating in USCF tournament players (which in of itself is far higher than the general population). 1800+ mathematically represents a significantly smaller subset

But of course you're entitled to call people shit with a timestamped picture of your FIDE CM title or better

So at it’as current state AI uses more CPU time (during training) costs more money (tensor) uses more memory than something like SF and will never be able to beat SF when we let SF do its work. SF is kind of a brute force tho which makes it less elegant.

But deciding which variations I can disregard is also intuition so I will blunder.

This guy gets it

This. A lot of "strategy" games just boil down to memorizing moves someone invented years before you, thanks to their finite mathematical properties.

Humans are poor at hard memorization though.
That's why more effective members use overarching concepts, patterns or themes for making choices.

>Take a game without randomness: since it has a finite number of states
its trivial for a game to have more states than you could ever possibly calculate in the lifetime of the universe

Pattern recognition is very useful in deciding which variations to play.

Afaik leela trains playing by itself like Alpha Go Zero, LC0 distributed community runs on regular OpenCL GPUs. SF brute force depth only goes so far even with enough cores, some of the leela's wins are nice positional plays blindsided SF's calculations. But yes, it currently loses most of its plays against Stockfish.

And retrograde analysis.

But not definitive. There are many such positions where the preferred process does not provide the correct solution.

>it’s not computable yet
It's computable enough that it makes no reasonable difference. AlphaZero can just go on playing against itself for another several decades until every sensible line has been analyzed. It won't be hard-solved but the unsolved parts will be so preposterous that just by trying to reach them you'd lose by move 20

No, humans improve with help of computers now, meaning it would be hard for humans to innovate new lines that computers don't know about.

>Chess is becoming a memory game
Chess was always built on pre-existing theory. Lasker and Morphy weren't making shit up on the fly back in the 19th century. That doesn't mean you win by memorizing. If that was the case, humans would have a chance against computers.

I'd say that Morphy if anything had an amazing implicit understanding and derivation of theory even if little explicit theory existed. Nimzovich laid the groundwork for the majority of formal theory which we use today.

>tfw I really hate playing against engines but I don’t want to lose ELO by playing people for fun.

What is the solution bros? Sometimes I just want to play for fun.

You play casual games? Like what you're forced to do when lichess bans you for cheating?

Play one game with an engine for every game you play for fun.

Look for local casual chess groups IE old men at parks and coffee shops.
I play with their pieces every Sunday personally.

Don't fear your rating. You should look at it only as a measure of your imporvement over time.

>so 99.999% is getting its ass whooped 100% of the time by that other 0.001%. Cool game bro
Is there any game where 99.999% is losing all the time to the top 0.001%? It's a truism that people who are extremely good at something will beat people who are not.