can magnus carlson cross the 3000 elo barrier

Wow Estragon, you really think Carlsen could reach 3000? With the current strength structure draws and loses are a major point loss for him...and he certainly can't win every game he plays. This is why I argued that more players above 2800 would be needed for him to accomplish this feat. What is he winning at now, about 70%. The problem I see is that as he increases his distance from the field so too will he be loosing more points for draws and loses.

Carlsen had a 3002 TPR in Nanjing 2009, after scoring 80% in one of the strongest tournaments of the year, and he had a 2994 TPR after scoring more than 80% in London 2012. He also scored 80% in Zurich 2014, which was enough to reach a 3027 TPR. Those are the three highest TPRs of his career. But scoring 80% and higher in such fields every single time is just impossible.

3000 will get 2.5 for winning 2800, which means he will need to score 75% against 2800 and impressive 85% against 2700 which is hard to believe...

The idea isn't that he has 3000 strength but can't reach the rating because of weak opponents. He doesn't have 3000

Here's a hypothetical scenario where Carlsen scores 8/10 in the next 19 tournaments he enters. We'll assume his average opponent is set at 2775, which is about right for the Nakamura's, Caruana's and Karjakin's of the chess world.

Initially, 8/10 would bring him big Elo gains against this field, but the rate of return soon starts dropping:

 Based on this, I can't see Carlsen ever getting much above 2950.

Looking at the last years of the 1990s, Kasparov scored 67% against opposition rated just below what was then top 10 level (2686). None of the hundreds of players on the list below reached 70%, so 80% today against 2775 (World #7) is a very high number:

http://members.aon.at/sfischl/1990b.txt

I think I remember hearing Kasparov say that "he should have found it" :)

I think it was him. All too easy to say as a spectator, though.

"This is why I argued that more players above 2800 would be needed for him to accomplish this feat."

A 3000 should be able to score a certain score against 2800s, and an even higher score against 2700s. That's what it means to be 3000. Someone who is truly 3000 should be able to get such an expected score against 2700s.

There could be some stylistic difference that makes it easier for him to play against 2800s than 2700s (keeping in mind against the 2800s he doesn't have to score as well), but then again it could just as easily be the other way around. You may say how hard it is to get so many wins against 2700s -- but it's even harder to do that against 2800s :) So I'm not sure why more 2800s would make Carlsen more confident about getting to 3000. It could slightly speed up the long run trend perhaps, but nothing more.

His draw with Karjakin (2772) cost him 1.5 points. He certainly can't win all his games, so loses and draws will cost him points. What is that ratio that he needs to maintain to continue to proceed forward? 2900 I can see, but the road to 3000 is too far ahead. He will need to win 80% of games playing 2700s...how possible is this? I think 70% is more likely. IMHO the road is too steep for him to reach 3000 given the current strength of his opponents.

Would it matter in this discussion if we assumed he either maintained his current strength with slight improvements over the years, or even a jump in his skills, and played to theage of, say, 45? I would be of the opinion that if he really sticks with it and doesn't A) Go insane like Fischer and B) Continues on for a decade or two either at his current strength or with improvements, then by 45 shouldn't he be perfectly capable of breaking 3000? To expect him to do it within the decade, however, seems to me to be asking too much of him.

I would agree with you generally NomadicKnight. I agree that he is still improving, and we have not yet seen the best of him. There are questions however about his opponents over the same period...since they play a part in him going forward or backward. There will be a number of them going downwards like Anand, Kramnik, and Gelfand but also younger players whom we can assume will reach 2800 like Naka, Caruana, and Karjakin. How much will they improve, and what impact will they have on him trying to get to 3000? I remember not too long ago Carlsen was one of that bunch. There are also questions about Magnus himself... now that he is the King of The Hill, how long will he continue to be hungary? Very difficult to remain on top with all your goals achieved. So many variables and hurdles in his path.

I agree. As much as I like Carlsen I'm not seeing him get to 3000. Not entirely impossible, but it would be extremely surprising.

With that said, it says a lot about Magnus's skill that we can even take the question seriously! We have to actually use arguments rather than just laugh it off as incomprehensible.

Can Magnus cross 3000?

Not if he cannot find a way to draw this game versus Caruana, no.

He's doing terrible!

Maybe he'll pull yet another swindle :p

Wouldn't people agree the position he has right now is still better than the one where he was losing against Naka?

His score now against Caruana (after this loss) will be +3 -3 =5

I hope people will now start to understand, after Carlsen lost two in a row, how hard eclipsing 2900 will be, and realize that 3000 is out of reach...

Carlsen gets ~1.5 rating points for his wins, loses points when he draws, and drops ~7.5 points for losses; thus, to gain ground he has to win more than 5 times as many games as he loses (not even counting the slow drain of draws).  This is what happens when your ratings gap is that much higher than all your peers, it becomes harder and harder to climb.

After losing 15 rating points in these 2 losses, he'll have to win 10 times against 2700+ opposition just to get back to 2889 (his high water mark).

Years of wonderful results apparently people don't care about when there are two days of poor results (ignoring two other wins he has had so far, mind you). And after all of that, at the end of the day he hasn't even lost ten points from his monster rating at the start of the tournament, even though it was his all time high! A "huge" loss of 8 points every few years... not exactly a roadblock for moving onto 2900.

Just amazing the conclusions people can jump to when they have an agenda.

That said, I'm not denying that making that last jump to 2900 isn't necessarily a trivial matter. Little disasters can hurt Magnus quite a bit, although it's not clear how often they will happen given how much of an anomaly his +0 score is. Fitting in a few more good results like he has been doing is hardly beyond Magnus, especially if he continues to improve.

2700chess disagrees with you.  Carlsen has lost 15 points since April 21st, 6.8 in his most recent loss (look at the "highest ratings ever" leaderboard on the right, not the main listing).

You have actually agreed that 2900 is not trivial, and I will assume you know 3000 is well out of reach for possibly decades...so we're actually in agreement.  You characterized me as naysaying Carlsen, but I did no such thing.  I am trying to eradicate ignorance about how the ELO system works.

"After losing 15 rating points in these 2 losses, he'll have to win 10 times against 2700+ opposition just to get back to 2889 (his high water mark)"

Is that really true? Even healthy +3, +4 results tend to give Magnus a decent rating boost. I'm not sure where that math comes from. Hell, I think he even gained one whole rating point by beating that 2500 during some club event a few weeks ago.

If Chuck Norris can't stop Magnis Carlsin reaching 2900, then so much for 2 defeats in a row. He's a choo-choo train.

btickler: Yes I know, regarding live ratings, his highest occurred during this tournament. My point is that as a whole, if the tournament just ended here, he could say his result was merely inferior, hardly a disaster. To have a tournament like this will make his rating worse off than if he didn't play in it, sure, but if he doesn't have too many of these results he should be in pretty good shape.

That is why I considered his 2881 rating more relevant to the discussion.