A 3000 could easily beat a 2000, but could a 4000 easily beat a 3000?

I don't understand this point of no rating being higher than 4000 either. Chess engines are already 3500-3600 and they make TONS of mistakes, can't solve puzzles like the one I posted, and do lose sometimes. I could easily see the perfect chess player being rated 1,000,000 or higher.

Yes, @EndgameStudier, your confusion lies in your lack of understanding of chess. Chess isn't just a draw, it has a very wide draw margin. A player can be up two knights and still not have enough to force the win. 

Take a look at the chess games of the best players. They all have individual games that are close to perfect, and that's with opponents trying to take them out of their preparation and into unknown territory. A 2800 will draw at least 1 in 50 games against a perfect opponent, especially if half of those games are with White.

Chess Engines cannot even solve this 15 move problem! So much for 3500 LMAO. A 2200 could prolly solve this in 5 minutes. If a chess engines can calculate millions of positions, don't you think this 15 move line would have came up?

I was talking a PERFECTLY PLAYED game, where both players make the best move every move until the game ends. Most assume chess is probably a draw, but we don't know this for sure. Although having an engine play another engine with alternating colors each time would solve this problem.

@Endgamestudier, I strongly suggest you study more endgames. You will find that there are more ways to draw a game than are dreamed of in your philosophy. 

LoL I do study endgames, but NO ONE knows if chess is a theoretically drawn game. Checkers was proved to be a draw using 11 supercomputers over an 18 year period! In chess, there are more possible games than their are atoms in the universe. If you assigned one atom per chess game, you would need over 10^50 UNIVERSES! Chess has not yet been proven to be draw, but maybe when they come out with quantum computers!

#181

@endgamestudier.

Simple.

I don't think that means what you think it means.

The logic behind it is simple. Keep triangulating the king around until black runs out of pawn moves and black is forced to move a knight and get mated. It's really only a 5 move puzzle. Think of it this way, if the king was a thousand squares away instead of only a few, would the puzzle be harder?

#185

@endgameStudier

1,000,000 or higher?!?!??

I disagree. Chess engines can solve something like 90% of chess positions at rating 3500. Maybe more.

The puzzle you posted is known about simply because it cannot be solved by an engine. That is why it is special. If engines were as bad as you seem to think they are a puzzle would not be noteworthy for being unsolvable.

I also ask when was the last time an engine tried that puzzle? What rating did that engine have? Has that alphachesszero engine tried it? Could it solve it given enough time? Are you sure this puzzle was tested in every engine?

You seem to think that puzzle would require an engine 30 times stronger then 3500 to solve it...

I mean if these engines can solve the very very very vast majority of all chess games and chess positions and are still learning and Improving, rated 3500, I highly doubt it will take an engine neatly 300 times stronger then the current (1,000,000 divided by 3500) ones to crack the rest of chess puzzles.

#192

True endgame, true. My mistake.

If a chess engine couldn't even solve that 15 move one I posted after, how strong does it have to be to solve a 70 move one? I put that into stockfish 9, which is rated 3390, and gave it 5 minutes to analyze, NOTHING. It doesn't even make sense, if it calculates million sof moves/second, even it didn't understand the "logic", it would have came across that 15 move line eventually.

Another question I have is when engines have multiple moves to checkmate, how does it decide which one? Like this position:

White has exactly 100 ways to checkmate, which move would the engine pick. I'm guessing Qxh8.

I agree, and we have the 50 move rule, and also threefold repetition that will make it hard to play 200 moves + games with a win.

And if the 4500 superengine plays open enginetournaments she will meet new and fresh 3000s giving her a draw now and then, which will cost maybe 4 ratingpoints every time (i donot know how much a draw do cost, 4 points is only a guess).

Well engines can calculate hundreds of moves ahead, so the 50 move rule isn't useful here. If players could see a hundred moves ahead, the 50 move rule wouldn't exist, because instead of just random moves, players would actually be winning those hard endgames! Engines can see those mate in 78s...etc, so the 50 move rule isn't necessary anyway.

Rating is relative. That's what makes this complicated here. If chess turns out to be a theoretical win for white and not a draw, then everytime tow PERFECT engines play, white will winning, costing the other rating points even tho there is no way she could have won anyway.

 I wish Anish Giri, one of the best ever to save a draw, could answer here.  He is only ca 2800 rated, and holds draws against  much, Or lets give Elias Hulleberg Sidali a try, when he was ca 1500 or 1700 he did draw an IM in a longchess tournament game, and also drew seveal other higher rated players. My guess is that a 2800 can save the draw with white at leat in 2 out of ten games regardless of their opponents strenght if they fight for draw from move one and do not overpush.

This, and the puzzles I posted that the engines couldn't solve, proves engines could still be greatly improved.

If you were talking about Go, which has a winner in almost every game, then the discussion of ever-increasing ratings would make sense. But in the context of chess, the conversation runs into the rules of the game. 

 

If you ignore the nature of chess, it's easy to create a scenario with almost any Elo rating.

What rule of chess am I ignoring? The 50 move rule doesn't apply when it comes to engines.