2200 vs 2700

I would expect more upsets in the 1400-1900 range, if only because many players in this range are improvers and their rating may be lagging with their actual strength (especially for young people). Of course, the higher # of blunders also accounts for some randomization.

@chessmicky : the score is probably closer to 92-93% when you remove simul games (can't do that automatically though, so just a rough guess). Expected % according to ratings is around 95%.

I used the (excellent) ChessTempo database for this search.

Basic Chess Endings and Fundamental Chess Endings.  Both are critically acclaimed especially the revision by Pal Benko (BCE).  However, there are still errors in it but given that Fine was dreaming with his "unrealistic" deadline to get such a massive work published very soon errors were to be expected.  Keres book is thinner but is said to contain fewer errors. 

Hmm, what do you mean tough player?  You work hard at the board (calculation) and have a good attitude (don't give up, or push hard for a win)?

Thanks for the explanation, I'll check them out if I find motivation for training again :)

I can think of several 2200-2300 players who do not have much chess understanding, but instead very good calculation, which leads to a very low (tactical) error rate and the ability to defend inferior middlegames and endgames against stronger opponents really well. If knowledge+understanding were everything in chess, child prodigies would have no chance against grandmasters, and the current world champion would be the winner of the match Kramnik-Ivanchuk.

Calculation is the most basic skill in chess as it determines who stands better at the end of either a defensive or attacking task and checks to see if a certain line fails due to tactical reasons.  Can never be too good at it although you have to know the elements of a position to understand why a certain line is good.  If one doesn't understand activity or initiative they could incorrectly conclude that since the opponent doesn't have isolated doubled pawns he must be better. 

Most excellent data, case closed. 2200 players are as fish to 2700 players.

Ok, I see.

Here's a different question I've been thinking about recently.

Some moves I consider quickly because I've seen them played before.  Maybe something like Qc7 or Re1.  Now I'm noticing many good non-tactical moves often involve some very simple logic.  e.g. maximizing mobility, making sure the piece can't be attacked / forced to move later, choosing a less vulnerable pawn formation. Something simple like this.

And yet I see some knowledgeable players make mistakes in this way.

Calculation is very important skill for sure.  As a lesser skill I'm wondering if some naturally strong players weigh moves this way.  And sure, when it's a natural move we all find it too, but sometimes it's a very odd move or plan but they find it easily because it's also very logical.

Yes. But the 2200 rated player is far more likely to be able to capitalize on far smaller mistakes. An extra pawn can be enough to beat the 2700 guy. 

I am an FM, and I have played against a number of guys between 2580 and 2700. My score is rather abysmal, I drew one and lost the remaining 4 or 5 games. Sometimes hard-fought, sometimes without any chance. But I'm not very "tough", always been much better at killing off the 2100-2250 crowd than at holding myself against IMs and GMs.

This is the sort of beating a gap of 200 rating points can lead to:

http://www.chessgames.com/perl/chessgame?gid=1519694

What does that tell you?

"Toughness" is an interesting idea.

I wonder what you would say makes you able to beat 2100-2200s?  I guess if you don't consider yourself "tough" then it's your theoretical knowledge or preparation?

I have a relatively good positional understanding, and a repertoire that supports it, and keeps weaker players from showing their strenghts. But this sort of detainment usually doesn't work against IMs and GMs. And when it comes to an open battle, they outcalculate me.

As someone said earlier in this thread, 500 points is 500 points it doesnt matter where it is. Its just math.There is nothing more or less likely about 2700 vs 2200, or 2200 vs 1700, or 1700 vs 1200.

The win percentage is what the ratings tell you, because thats how the rating is determined in the first place. Its really quite simple.

If the 2200 player was winning more often against the 2700 player, the rating would change to reflect that, and he would become a 2300 player.

Fantastic thread btw!!  what i can take away from this is no matter your skill level always look for an opportunity to perform tactics or set up a favorable endgame dynamic.  A mistake in not just losing a piece but rather missing an opportunity..

http://www.chess.com/forum/view/game-showcase/drawing-a-2700-gmas-black

Sorry, couldn't resist.

Wow, that's awesome, good job!

a 2700 would beat a 2200 in the same manner as a 2200 would beat 1700. 

also don't really think 2200s have a "profound" understanding of chess...... 2700? yes. 

I had the version revised by Benko ever since i was a kid. 

It seems like 500 points is 500 points but you have to understand that the higher you go the higher the diminishing returns.  A 50 point gap at the 2700+ level is huge whereas it's almost insignificant at club level.