I didnt even think of that. The ability difference is not the same in 4000 vs 3000. Its only 1.333 x the ability, where 3000 vs 2000 is 2.5 x.
That's not how it works.
9 out of 10 (or more) comments in this topic are completely stupid.
Think of it this way: There are about 50 legal moves in any position, so let's say a 1500 will see the 10th best moves, a 2000 the 5th best moves, a 2500 the 2nd best move, and a 3000 THE best move. This means any player above 3000 will most likely see the 1st or 2nd best move in a position. So a 4000 would see the best move, and so would a 3500. It doesn't matter if someone is rated 5000 or 1000000, they will make the best moves most of the time and result in draws.
I agree with EndgameStudier. After 3500 rating, they're all going to play the best move, so there will be no butts being whipped at the boards. Mostly draws and probably mostly pre-arranged agreed upon draws, so the competitive aspects will be lost.
I agree with EndgameStudier. After 3500 rating, they're all going to play the best move, so there will be no butts being whipped at the boards. Mostly draws and probably mostly pre-arranged agreed upon draws, so the competitive aspects will be lost.
Yes, but that's assuming chess is a draw when played perfectly. What if White Wins if both players play perfectly, then all the white players would win.
Think of it this way: There are about 50 legal moves in any position, so let's say a 1500 will see the 10th best moves, a 2000 the 5th best moves, a 2500 the 2nd best move, and a 3000 THE best move. This means any player above 3000 will most likely see the 1st or 2nd best move in a position. So a 4000 would see the best move, and so would a 3500. It doesn't matter if someone is rated 5000 or 1000000, they will make the best moves most of the time and result in draws.
So, you made up some statistics and then used invalid reasoning on them. How reliable do you think the results are?
From the point of very of a perfect player (the best judge) there is no order of moves in a position. Rather there are three (or fewer) classes of moves.: those that win, draw or lose against perfect defense.
Games are lost by the player who is the last one to play a move that loses when there is a move available that doesn't. This move could be the "second best move" if there was only move that had a better theoretical result. No players including the strongest computers are invulnerable to making such blunders: they just get less vulnerable to doing so. In addition, strong players provide more opportunities for their opponents to blunder and don't let them off the hook often when they do.
Think of it this way: There are about 50 legal moves in any position, so let's say a 1500 will see the 10th best moves, a 2000 the 5th best moves, a 2500 the 2nd best move, and a 3000 THE best move. This means any player above 3000 will most likely see the 1st or 2nd best move in a position. So a 4000 would see the best move, and so would a 3500. It doesn't matter if someone is rated 5000 or 1000000, they will make the best moves most of the time and result in draws.
So, you made up some statistics and then used invalid reasoning on them. How reliable do you think the results are?
From the point of very of a perfect player (the best judge) there is no order of moves in a position. Rather there are three (or fewer) classes of moves.: those that win, draw or lose against perfect defense.
Games are lost by the player who is the last one to play a move that loses when there is a move available that doesn't. This move could be the "second best move" if there was only move that had a better theoretical result. No players including the strongest computers are invulnerable to making such blunders: they just get less vulnerable to doing so. In addition, strong players provide more opportunities for their opponents to blunder and don't let them off the hook often when they do.
Are you seriously gonna attack me on this thread too? You get my point. There are a few dozen moves in any position, and better players will play the better moves, and after a certain level of ability, players will make the best move, regardless of whether the players are 10,000 rated or 1000000000000 rated.
Think of it this way: There are about 50 legal moves in any position less than that, so let's say a 1500 will see the 10th best moves no no no, that's not how it works. Competant players well below master can often find the best move. Chess is hard because it only takes 1 bad move to lose. , a 2000 the 5th best moves, a 2500 the 2nd best move, and a 3000 THE best move. This means any player above 3000 will most likely see the 1st or 2nd best move in a position. I, at a sub-master level will "most likely see" the 1st or 2nd best move in most positions. So a 4000 would see the best move, and so would a 3500. It doesn't matter if someone is rated 5000 or 1000000, they will make the best moves most of the time and result in draws. Elo doesn't top out at 3000. That's silly for even the most basic error check. E.g. Carlsen is 2800 and will lose 10 out of 10 against an engine (so that's at least a 400 points difference), who lost a match against AZ.
We were just talking about driving in the other thread, Elroch. If you slam into a brick wall at 1000 mph, it doesn't matter if you are going 1000 or 2000 or 3500 mph..etc, you will still die. Maybe going 60 vs 70 mph will make a difference, but over 1000 mph it's all the same; Same logic here.
And since ratings are relative, if you're claiming 3000 is at (or very near) the top. then 1000000 wouldn't exist.
We were just talking about driving in the other thread, Elroch. If you slam into a brick wall at 1000 mph, it doesn't matter if you are going 1000 or 2000 or 3500 mph..etc, you will still die. Maybe going 60 vs 70 mph will make a difference, but over 1000 mph it's all the same; Same logic here.
In other words you don't know how the rating system works, so you resort of analogies.
Think of it this way: There are about 50 legal moves in any position less than that, so let's say a 1500 will see the 10th best moves no no no, that's not how it works. Competant players well below master can often find the best move. Chess is hard because it only takes 1 bad move to lose. Wrong, many times tow moves that look just as good are often very different, and many best moves are subtle. , a 2000 the 5th best moves, a 2500 the 2nd best move, and a 3000 THE best move. This means any player above 3000 will most likely see the 1st or 2nd best move in a position. I, at a sub-master level will "most likely see" the 1st or 2nd best move in most positions. Maybe in a mate in 3 position, or in the opening, but in a complicated position, you probably won't. Don't get me wrong, you will probably make really good moves, but not the BEST move. So a 4000 would see the best move, and so would a 3500. It doesn't matter if someone is rated 5000 or 1000000, they will make the best moves most of the time and result in draws. Elo doesn't top out at 3000. That's silly for even the most basic error check. E.g. Carlsen is 2800 and will lose 10 out of 10 against an engine (so that's at least a 400 points difference), who lost a match against AZ.
Ok, let's say 10,000 for the sake of argument.
We were just talking about driving in the other thread, Elroch. If you slam into a brick wall at 1000 mph, it doesn't matter if you are going 1000 or 2000 or 3500 mph..etc, you will still die. Maybe going 60 vs 70 mph will make a difference, but over 1000 mph it's all the same; Same logic here.
In other words you don't know how the rating system works, so you resort of analogies.
I know how the rating system works, and I know a 1400 will beat a 1000 much more easily than a 2800 will beat a 2400. If we can't even agree on this then I don't know what to say.
And since ratings are relative, if you're claiming 3000 is at (or very near) the top. then 1000000 wouldn't exist.
Did I say 3000 was the top? Did I say ANY rating was the top? There are a finite number of chess games, but an infinite # of integers, so above a certain rating, both players would play a PERFECT game.
And in addition, rating is based on RESULTS, not how good of a game was played. If I'm a 2000 and I keep beating a 1200 over and over again, I will keep gaining at least ONE point, but that point doesn't mean anything. Correct me if I'm wrong, but I think a rating always changes at least 1 point with a win or loss.
Putting the other crap aside, ok, lets say 10000 Elo = perfect player.
Then 10,000 will win nearly all games against 9000... otherwise the rating 10,000 wouldn't exist.
Putting the other crap aside, ok, lets say 10000 Elo = perfect player.
Then 10,000 will win nearly all games against 9000... otherwise the rating 10,000 wouldn't exist.
Not necessarily, even if both players play a perfect game, that doesn't mean it will be a draw. Chess played perfectly might be a win for white, or even black for all we know, so perfect players would result in the same color winning every time. We cannot assume that a perfectly played chess game does not end up in a win or loss for one side.
When chess.com was still new, yes, ratings always changed at least 1 point. That's how some NM got to 4200 or some silly rating like that. Chessnetwork IIRC.
But no, gaining 1 point per win minimum is not how Elo or Glicko works. And chess.com has fixed that now.
I believe someone posted the Elo rating formula earlier. Plug in numbers and you'll see.
Long story short, it's always some number between 0 and 1 multiplied by the K factor, which is usually 20.
So for example .10 x 20 = 2. So the rating would change by 2 points.
I don't need to tell you something like 0.0001 x 20 is not going to change it by 1 
Putting the other crap aside, ok, lets say 10000 Elo = perfect player.
Then 10,000 will win nearly all games against 9000... otherwise the rating 10,000 wouldn't exist.
Not necessarily, even if both players play a perfect game, that doesn't mean it will be a draw. Chess played perfectly might be a win for white, or even black for all we know, so perfect players would result in the same color winning every time.
You keep ignoring the fact that ratings are relative to each other...by definition a rating difference of 1000 points, regardless of where on the scale that difference occurs, means that the higher rated player will win 99.7% of the time. That's the way the ELO system works, if two players are drawing almost all of their games they cannot be 1000 points apart on the scale.
I didnt even think of that. The ability difference is not the same in 4000 vs 3000. Its only 1.333 x the ability, where 3000 vs 2000 is 2.5 x.