Could a 2000 rated player beat Magnus Carlsen?

Sleepyness can be a problem. I once made a blunding-blunder, playing 10 minutegame here. My finger was on the scrollpad, i fell asleep, and suddendly a woke up because of a "pling", my finger had moved and put my knight in a square where it was taken for free.

lol.that was strange.or make Magnus carlsen tired wit hjet lag and play him

I'm quite familiar with how ratings work.  Are you?  Elo presumes a normal distribution and uses parametric statistics for its calculations.  Parametric statistics are predicated on interval or rational scaling and chess does not provide that.  So it is a leap of faith to presume that the curve is not skewed.

Nominal and ordinal scaling require non-parametric statistics.  The downside is that those tests like the Spearman Rho and Chai Square do not provide the power of parametric statistics like the T tests and especially do not provide the power of someting like an ANOVA or ANCOVA which depend exclusively on rational scaling.

Chess games are on ordinal scales and don't merit Elo's parametric methods.

An interval or rational scale measure the same thing over time.  A fever of 103F doesn't have an inflationary value where that becomes a normal temperature 20 years later.

Even if one granted a normal distribution it is still obvious that the _qualitative_ difference between a 1000 and 1800 and 2000 and 2800 is different.  An 1800 is going to have some knowledge of openings, tactics and perhaps end games.  A 1000 is probably going to know how to make the moves.  When An 1800 is playing out a number of standard moves from book she is essentialy playing out grandmaster and computer analysis and thus playing well beyond her strength in that part of the game.  When we get "out of book" we fall back on our own resources.

The _qualitative_ difference in the game between a 2000 and 2800 is likely going to show the 2000 playing relatively strong moves in certain parts of the game while playing like a duffer in others.  A 1000 rated player is going to suck in all aspects of the game-- a qualitative difference.

Claiming a standard distribution doesn't make it so.  Modifying equations to artificially create a standard distribution from current scores doesn't make that a true equation.  It has descriptive power but predictive power is far more limited.

I think it's important to stipulate "can a 2000 beat a 2800 in a one-on-one game at standard tournament time controls."  That's what the elo is measuring.

The chance that will happen is almost 0.1%, which means out of 70 games 2000 has 50/50 chances to score any points

So, the higher the rating , the harder it is to gain new points.

i agree

very true

I think 2000 players are very strong,and guess that their best games are played at 2500+ strenght. Carlsens worst games are probably played at below 2500 strenght.

Carlsens worst games are played at like 2650

I wish it was true, but I know it isn't.

EDIT : quote truncated to the relevant part.

The question has already been answered. Carlsen has lost to a +2000 rated player in a simul. 

My Nephew at 1850 Fide have been playing equal with an IM. Therefore I guess that Fide- 2000 players are playing their best games at GM strenght.  I think many of the young talents are on the way up, and they can play very strong, but they dont have a total chessknowledge , like the GM`s, so they  loose a lot of games too,when playing strong players.

On my club we had a tournament when two players 1086 and 1900 met. The low rated player managed to take the queen and was even a piece up. But because his lack of time he made 3 bad moves in a row. The 1900 player won. but it was close.

I drew with Tal in a simul in '88. He was just lucky...

The Olympic games shows that difference in rating means less than I did expect. 

I do play at a lot lower level, but what I see in my own games, is that I myself am my worst enemy, and that losses most often comes because of my own mistakes.  I guess this to a certain degree goes for the GM´s too.

Several years ago I was reading a book and don't remember the title but from what I recall Bobby Fischer never lost a single game to an International Master after he won his GM title.  I don't believe he was ever held to a single draw against anyone less than a GM.  That would be a good comparison.

Bobby Fischer's score of 19-0 against world class players will probably never be beaten.

oh that

i remember reading a book like that

Magnus lost today to a player 200 points below him. Admittedly, he tried a strange double pawn sacrifice for later compensation, which his strong GM opponent was able to deal with. It's doubtful a 2000 player could take this opportunity and go with it. Perhaps a 2000 that is highly underrated and on his way to becoming a GM.