Imagine Magnus Carlsen being able to score 66% against his old self all the time.
can magnus carlson cross the 3000 elo barrier
Sort:
scaccomato wrote:
ELO is not an exact moving micro measure of a chess players ability at playing chess. There are many variables involved in a chess player's performance in every game, such as adequate rest, domestic matters, health, to name just a few that ELO is unable to factor in. Hence, slight differences in rating say 2890 v.s. 3010 is really almost meaningless. Does not mean that the former cannot beat the latter, as we could say with certainty that heat measuring 110 degrees Farenheight is always hotter than one below it. ELO is not a speedometer, or tachchometer.
On the contrary...at the end of the scale without anyone above you, a 120 point rating diffential is gigantic. Even though the winning percentages don't change, the lack of anyone above you to win significant ratings points from makes a huge difference. So, a 1600 moving to 1720? Not that difficult. Knock off a few 1800s and you are on your way. Moving from 2800 to 2920? Incredibly difficult. In an Elo ratings pool, the outliers at the top and bottom end are naturally drawn back into the fold like a blob monster from a 1950's movie...it becomes harder and harder to break away, and at 300 (400? have to double check) rating points above or below the pool, it becomes mathematically impossible.
So, the fact that a 1600 player can sometimes upset a 2000 on a bad day is not at all germane to the argument of Carlsen trying to reach 2900 or 3000. The factors you listed affect performance ratings in single tournaments, not overall ratings in the long term.
Think of 3000 rating for the moment like the speed of light...the laws of physics start to behave differently when you approach the barrier. Carlsen cannot reach 3000 with his peers at 2750-2830...his acceleration is overcome by infinite mass ;).
"Carlsen cannot reach 3000 with his peers at 2750-2830...his acceleration is overcome by infinite mass ;)."
You would agree, though, that even if his opponents were "genuinely" 2900 (that is, both have a 2900 rating and have the strength that would today get your strength to 2900 -- e.g., similar to how well Magnus is playing now), then Magnus's situation wouldn't be any better.
In other words, when you say his opponents need a higher rating, you mean only a higher rating, and not actually a higher playing strength. So you mean only that, say, Anand would need to stay at the same strength but have his rating inflated so that his rating was 2900 or something.
Elubas wrote:
"Carlsen cannot reach 3000 with his peers at 2750-2830...his acceleration is overcome by infinite mass ;)."
You would agree, though, that even if his opponents were "genuinely" 2900 (that is, both have a 2900 rating and have the strength that would today get your strength to 2900 -- e.g., similar to how well Magnus is playing now), then Magnus's situation wouldn't be any better.
In other words, when you say his opponents need a higher rating, you mean only a higher rating, and not actually a higher playing strength. So you mean only that, say, Anand would need to stay at the same strength but have his rating inflated so that his rating was 2900 or something.
Actual playing strength is irrelevant in this particular context. Have you ever seen Moneyball? It doesn't matter if Carlsen and all his opponents play after taking 15 jello shots each from this day forward, all that matters is the pool he is in and his win/draw/loss percentages, i.e. his relative playing strength vs. the pool.
You could take all the monkeys typing Shakespeare and swap them in for the top GMs and name one Carlsen...if the Carlsen monkey wants to break 3000, he has to be the same X% better than the other monkeys. You could change the game of chess to tiddlywinks and call it chess...
"It doesn't matter if he's a actually a good hitter or not...his onbase percentage is .322 and that's all that matters. He gets on base."
So, yes, I mean that his peers need higher ratings, not to be better at chess :).
Ok. Because if Anand was 2900 because he just won the next 50 games he played against 2700s (whose strength remained the same), and then played Carlsen in a match, Carlsen's chances of becoming 3000 would not increase (or decrease). Magnus would get more points for each win, then give them back when he lost. This is assuming of course that Anand moved up from the high 2700s to 2900 because he is now a more difficult opponent to beat, and will in general have a higher win percentage against Carlsen.
So I wouldn't say playing strength is irrelevant, since in this example Anand's increased playing strength did in fact give Magnus a 2900 opponent (although as I said this won't help Magnus reach 3000), but of course relative playing strength is the star of the show.
It's only because we know of the kinds of players a 2800 today beats that we attatch so much meaning to the number. With the BCF system, I think even a 200 rating would be pretty godlike.
So I think we agree on the importance of winning percentages and such (although disagree on whether rating inflation exists at this time).
Is it so that there have been a ratinginflation of ca 100 points the last 40 years?
Can we ecpect further inflation?
Maybe Magnus can gain some strenght and some inflation and make 2900 in 40 years? But that is difficult, because there are very sharp players coming up now, that will be very strong 10 years from now. My club is stuffed with talents that at least have IM-possibilities, and maybe super GM.
Lets say that ratingpoints are harder and harder to eat, the higher you are. If Magnus are eating the coming years 15, 11,10,9,9,9,8,8,8,8,8,7,7,7 points,then he is getting close to 3000 in ca 15 years.
There is a huge risk Magnus is facing . The ladys. He might get married. What happens in a marriage, can show in the ratingnumbers.
Estragon wrote:
The question of "inflation" is a different one, which is not possible to measure. The best players of each era are stronger than those before, with a few shining exceptions. And each builds off the refinements of those before. Chess has not stopped improving in quality at the highest levels.
The top 20 or so players of today are not there because of inflation. If you judged their moves against the strong engines, they would do better than players 20 years ago, or 40 or 75.
The idea that Carlsen "can't" achieve 3000 for mathematical reasons is not correct. It assumes that the others in the top 20 and top 50 will not continue to improve their games and ratings. That's a false assumption.
It is also not the case that the top rated player cannot gain rating points. It is more difficult at that level because of the low "k" value, it is not possible to move up quickly once you are on top. But Carlsen took over first place for the last time in July 2011 at 2821. In the 22 months since then, he has gained 60 points.
It's almost impossible to gain 30 points a year at his level. But he's been doing it. There is nothing magical about 2882 that makes it less possible to advance than 2821 does. It is a matter of results of course.
In 2011, Carlsen was just pushing ahead of Anand, Kramnik, Topalov, and Aronian. Now he is clearly the superior of all of them.
He's 23. The history tells us that the great masters tend to hit their peak at 32. He is getting better.
Carlsen had a bad tournament at the Super-GM Gashimov Memorial, lost two games in a row, bungled a win into a draw, and still managed to win the tournament with +4 in ten games.
It could easily take five years, but it is not at all inconceivable that Carlsen could reach 3000 if his progress continues.
For Carlsen to reach 3000 in 5 years, he would have to gain 120 rating points. That would be almost linear with his 60 point gain in the past 2 years. With Aronian at 2830, the top of the ratings pool is like an asymptote for Carlsen, with 3130 being effectively infinity, in terms of achievable winning percentage vs. the pool. Postulating an easy 3000 rating from 2900 is more like hiking up Mt. Shasta, and then claiming you can climb K2 because it's only twice as tall.
If Carlsen makes 3000, it will probably be because of changes by FIDE, or slow inflation of the pool.
How many points will he drop if or when he looses for a 2350 player?
-9.2
It is inconceivable that Carlsen can reach 3000. He hadn't had a single tournament performance that would indicate that it's achievable.
He has had 3000 performance ratings in tournaments. But those were somewhat special, even for him. He'd have to be that brilliant every tournament he plays for years to come, and any time he messes up he will lose a lot of hard-earned progress.
You can see the difficulty outlined here: http://www.2700chess.com/glass.php?statistic=2
Carlsen had 8 good games, where he scored 6.5/8. If he'd only played these 2 games his rating would now be 2895. Alas, he lost 2, and therefore his rating actually went down overall!
| Points | Games | Rating |
| 6.5 | 8 | 12.3 |
| 0 | 2 | -13.5 |
It's a credit to Carlsen that he nearly always goes for the win, despite the effect a loss will have on his rating.
no, but he will break the 2900.
I think that there will always come strong players from below, and that he can loose for some underrated 2400 players on their way up. losses that costs a lot of ratingpoints.
I know a 17 year old talent that went from below 2000 to 2350 in a year, players in this category can gain strenght faster than they gain ratingpoints, and du serious ratingdamage to higher rated players, that not necessarily are stronger. I think Magnus is the strongest, but he too are taking a loss now and then.
Elubas wrote:
He has had 3000 performance ratings in tournaments. But those were somewhat special, even for him. He'd have to be that brilliant every tournament he plays for years to come, and any time he messes up he will lose a lot of hard-earned progress.
He wouldn't break 3000 with 3000 performances for the rest of his career.
Hence, slight differences in rating say 2890 v.s. 3010 is really almost meaningless. ...
Ummm....
A difference of 120 rating points means that the higher rated player is expected to win 2 out of 3 points. http://www.pradu.us/old/Nov27_2008/Buzz/elotable.html
Is that really almost meaningless? I suppose, from a rather nihilistic perspective, all ratings are meaningless. But a difference of 120 rating points is pretty significant.