A super Gm is what seperates the elite and the best !
What's the difference between a Super GM and a GM?

R_N_A: to give an idea of what kind of answer you have in mind, could you describe to us the difference between an expert and a master? Or between a 1600 and an 1800?
In general, I think the difference really is that the higher rated player has the same stuff, only more of it. More patterns. More opening theory. Better endgame technique. Better visualization and calculation. Fewer hanging pieces. I strongly suspect the same is true of any 200 point rating difference, regardless of whether you're comparing 1500 vs 1700 or 2500 vs 2700.
At any rating level X, you will naturally have some players who are above average at opening theory but below average at spotting tactics, or something like that. But for any given area of chess knowledge/ability, the average player of rating X+200 will have more of it than the average player of rating X. You can't narrow things down to "the difference between 1600 and 1800 is in opening theory" or "the difference between 1800 and 2000 is in endgame knowledge" or "the difference between 2000 and 2200 is in piece activity".
Best answer
Thread closed, unless the rest of you useless members want to keep on discussing underpants.

R_N_A: to give an idea of what kind of answer you have in mind, could you describe to us the difference between an expert and a master? Or between a 1600 and an 1800?
In general, I think the difference really is that the higher rated player has the same stuff, only more of it. More patterns. More opening theory. Better endgame technique. Better visualization and calculation. Fewer hanging pieces. I strongly suspect the same is true of any 200 point rating difference, regardless of whether you're comparing 1500 vs 1700 or 2500 vs 2700.
At any rating level X, you will naturally have some players who are above average at opening theory but below average at spotting tactics, or something like that. But for any given area of chess knowledge/ability, the average player of rating X+200 will have more of it than the average player of rating X. You can't narrow things down to "the difference between 1600 and 1800 is in opening theory" or "the difference between 1800 and 2000 is in endgame knowledge" or "the difference between 2000 and 2200 is in piece activity".
Best answer
Thread closed, unless the rest of you useless members want to keep on discussing underpants.
Like you can decide. Sheesh.
Bramblyspam does seem to have a pretty good answer, but nevertheless... how can he really know? It could vary depending on the level of play. For example, maybe there are certain kinds of concepts in chess that if you grasp it, you will do way better as a result, and for whatever reason, chess players below a certain rating x just typically don't feel like learning it -- maybe because it's difficult, or too deep. And maybe it just takes something exceptional to grasp it sufficiently.
For example, maybe all 2000 players have very good tactics (even though some even better than others), but there is some kind of efficient way of thinking that if learned, would get them to 2200, even with a similar amount of tactical knowlege. Maybe it's because they just can't afford to get bad positions anymore -- even if they are outstanding at tactics, the tactics will favor them so little that they will still inevitably blunder in a bad position anyway.
So there can become some sort of wall, where, yes, the 2000 could keep improving tactics, and maybe get marginally better, but will still inevitably blunder in bad positions out of the opening all the same because they are just so hard to play. In that case they would lack something exceptional that, say, 2200 players might have.
Such a wall is less likely to occur at, say, 1400 level, since most games are just decided by blunders at some point. Improving tactics will just proportionally improve your results almost.
Is any of this the case in my example? My guess would be no, that the truth is closer to what bramblyspam is saying. My point is it's pretty hard to know that. We don't know 100% the kinds of knowledge that are most emphasized by each rating level. Sure, we know that, on average, a higher rated player will be better in every facet compared to a lower rated one. But we don't know the average when it comes to how much better the stronger player tends to be in each category.
It might be that, for example, a 2200 is on average just a tiny bit better at tactics than a 2100, but much better at openings. The reason for that might be, once one's tactics reach a certain point, you pretty much won't blunder in easy-to-play positions, and so, the cause of losses will start to be more related to getting hard-to-play positions, where improving at the opening starts to be more rewarding, and thus, explains an average 2200's better results compared to an average 2100.

Super GM's know 5,000,000 patterns
GM's know 20,000 patterns
Interesting post. Is there a source for these numbers?
What?!? One number is 250 times bigger than the other! He was clearly joking, there is no source.
You could have just said no, but, whatever.

Best answer
Thread closed, unless the rest of you useless members want to keep on discussing underpants.
LOL that's the answer you're happy with? Obviously it's the correct answer, but any idiot could have given you that correct answer on page one. I know I could have, it's just the most basic of common sense. Nothing in that answer was specific to 2700s, he just described what being a better player than somebody else is like. I can't believe that's the answer you're happy with. I think you just got bored of the thread.

In think the most interesting answer is to examine the games between them. By doing so we can aquire some good understanding ourselves, and that will help our games.
Look at the games of GM Aleksey Dreev. I played trough the game versus Tanguy Ringoir. Game pops up when clicking the chequered icon.
It was an interesting game. This game was tactical.
http://www.xtraconchessopen.dk/index.php?did=2769&mid=94

Trolling is not necessarily bad. Good ideas can evolve, and we are free to post the best we can write.

The correct answer!
A Super GM CAN crush a GM
A GM CANNOT Crush a Super GM
Simple..
To beat a 2500 GM is as Alex pointed out not easy for anyone, especially not if the superGM plays black. A superGM usually draws lighter GMs with black, and if the SGMs gets to sharp and ambitious they can lose. SuperGMS are very good at not loosing. When GM Hammer became super he had 49 tournamentgames in a row without losses. I think Ivanchuk is taking slightly more risks, which results in less draws and more wins and losses than Hammer gets. Urkedal is a talent not less than Hammer. He was a few years ago Norwegian Champion, but has slowed down sligthly in his improvements because of too much dedication at his universitystudies. Urkedal has, except for economy, what it takes to become super. He has climbed to 2557 now, so the gap is closing.
"Johannes pushed hard with black but couldnt get anything, Shirov squeesed in the endgame, and after 54 moves he got desisive advantage." "It looked equal (in my very amateurish eyes) after 25moves, but from there Shirov slowly bulidt an advantage, starting with entering the 7th rank with his rook."
Yes. It is always interesting how a strong GM beats other masters.