Well, I don't know much about the Grunfeld, but Black could probably always win it back (say with 10...Qa5), if he wants to, and then White will have lost his strong pawn center.
One moment White has a c3/d4/e4 pawn center, restricting the Black pieces, always threatening to advance one of the center pawns driving Black's pieces away; then after d4xc5 I just see a bunch of weak isolated pawns on the queenside.
Scarblac's about right. 10.dxc5 and White is voluntarily giving up his centre. Effectively Black has achieved all of his opening aims without having to do anything.
As a Gruenfeld player myself, I must (sadly) confess that I've lost blitz games to players who decide to grab the pawn and hang on for dear life. I know a lot of older club level players who have decided that the correct way to play chess is to grab pawns. Quite honestly, if you look at the position carefully, you can pretty much win it back by force. The grunny is all about preparation... in the position you give there for example, I play 10. ... Bg4 (usually, there is also b6 and Qc7, both which I know well) and after 11. f3 Na5 white can again choose to grab a pawn with 12. Bxf7+ Rxf7+ 13. fxg4 ... some of the theory in these lines run nearly 30 moves deep. If the Karpov - Kasparov matches mean anything, black is fine here. Weird eh? :P
Anyways, as for grabbing the pawn, I'd suggest (and I don't suggest this for just about anything else) whipping out a database and rybka... and memorizing a few lines/ideas. You'll quickly figure out what the problem is.
d4xc5 should result in weak pawns for white. Black's reaction could be ...e5 to further exploit the d4 square. However, I have seen d4xc5 work nicely. In the following game, white puts a piece on d4, induces an exchange on d4, recaptures c3xd4, and gains an impressive phalanx of pawns.
You should better familiar yourself with the positional idea behind such maneuver (sacrificing a pawn for two weak pawns) so you can employ it whenever you get the chance.
Sometimes you don't need to know precisely how you can regain your sacrificed pawn. You give away one pawn to create two weak pawns. With such weak positions in many cases it is almost always difficult to defend any one of them.
And after one pawn down, your opponent is left with one weak pawn. And that is your advantage (your objective when you sacrifice a pawn).
It's been mentioned that white gives up his strong pawn center, but no one has yet pointed out that the bishop on g7 instantly becomes far and away the best minor piece on the board. Add this to the already impressive list of compensation that's been given in this thread.
d4xc5 should result in weak pawns for white. Black's reaction could be ...e5 to further exploit the d4 square. However, I have seen d4xc5 work nicely. In the following game, white puts a piece on d4, induces an exchange on d4, recaptures c3xd4, and gains an impressive phalanx of pawns.
Lol at Bd7!
Actually there are lots of different ways black can react to this. It depends upon when white takes the c5-pawn, which reaction is good.
Black could make it a real pawn sacrifice (or maybe even a double one) by answering dxc5 with ..b6!? at some point, aiming to open the a- and c-files for his rooks, pointing down at white's weak remaining queenside pawns. This can be played in conjunction with Na6/Nd7.
Ofcourse, Na6/Nd7 can be played without b6 as well.
And there's Qa5, as scarblac mentioned.
And I'm sure there are a few other ideas I'm overlooking at the moment.
The point is that, usually, if white plays dxc5, this gives black very good counterplay - it is almost impossible to hold on to the pawn white has just 'won'. dxc5 is in fact a typical computer move, because it is materialistic, but the computer has a horizon - it cannot see enough moves ahead to know that it is going to have to return the pawn later anyway with an inferior position.
The Grunfeld is based on play against the dark central squares and active piece play. In many variations Black must be willing to sacrifice a pawn for position or activity and initiative to avoid a positional disadvantage.
It takes a certain personality to play the Grunfeld, an aggressive person with a sharp tactical eye who is willing to trade a disadvantage for another advantage even when it is risky and complicated.
Yes, I took up the Grunfeld just over a year ago, and I was surprised by the sheer number of lines where Black sacrifices a pawn!
Alphastar: Certainly in the position by the OP, 10.dxc5 is not a 'computer move'. Computers these days are well able to evaluate that the capture is not beneficial to them.
Alphastar: Certainly in the position by the OP, 10.dxc5 is not a 'computer move'. Computers these days are well able to evaluate that the capture is not beneficial to them.
I disagree and already explained my position. Maybe you can give some arguments.
Anyways, as for grabbing the pawn, I'd suggest (and I don't suggest this for just about anything else) whipping out a database and rybka... and memorizing a few lines/ideas. You'll quickly figure out what the problem is.
If you are a Grunfeld player as you say, you should have been able to outline IDEAS rather than saying "memorize some lines". A good player does not memorize lines, but knows what are the plans in the opening and where he wants his pieces to be.
Alphastar: Certainly in the position by the OP, 10.dxc5 is not a 'computer move'. Computers these days are well able to evaluate that the capture is not beneficial to them.
I disagree and already explained my position. Maybe you can give some arguments.
I looked at the position at move 9 in the original post with Fritz8. At 11 ply, 9. dxc5 was 7th on the list with evaluation -0.03, the top two moves were 0.28 and 0.25. At 12 ply 9. dxc5 doesn't even make the top 10.
Fritz8 isn't exactly up to date software. But already by that time computers could see deep enough and use the info about the weak pawns to stay away from moves like 9. dxc5. State of the art software will see a little deeper and with improved evaluation -- no computer these days is going to make the mistake 9. dxc5.
Two decades ago, this is the kind of mistake computers would make.
Anyways, as for grabbing the pawn, I'd suggest (and I don't suggest this for just about anything else) whipping out a database and rybka... and memorizing a few lines/ideas. You'll quickly figure out what the problem is.
If you are a Grunfeld player as you say, you should have been able to outline IDEAS rather than saying "memorize some lines". A good player does not memorize lines, but knows what are the plans in the opening and where he wants his pieces to be.
Being a tactical line. The gruenfeld often requires pure memorization. 
Many of the ideas have already been covered. There's the c5 break, an e5 break, and an f5 break (in that order) ... keep those moves in mind. The idea is to destroy or comprimise the white center. Take a look at my blog "A small giant" for a positional look at the gruenfeld.
Anyways, as for grabbing the pawn, I'd suggest (and I don't suggest this for just about anything else) whipping out a database and rybka... and memorizing a few lines/ideas. You'll quickly figure out what the problem is.
If you are a Grunfeld player as you say, you should have been able to outline IDEAS rather than saying "memorize some lines". A good player does not memorize lines, but knows what are the plans in the opening and where he wants his pieces to be.
Being a tactical line. The gruenfeld often requires pure memorization. Many of the ideas have already been covered. There's the c5 break, an e5 break, and an f5 break (in that order) ... keep those moves in mind. The idea is to destroy or comprimise the white center. Take a look at my blog "A small giant" for a positional look at the gruenfeld.
I have the Dembo Grunfeld book, which is pretty good. is there a certain book you rec. to book up on it?
I am looking into the Grunfeld because I am a little tired of playing KID...but goodness, this opening has so much to be learned.
I am looking into the Grunfeld because I am a little tired of playing KID...but goodness, this opening has so much to be learned.
Does the Grunfeld really have more to learn than the King's Indian? It seems to me that the King's Indian has a lot more to learn, and that the Grunfeld is more straightforward, but perhaps I'm just missing something. (?)
You're missing something, Biarien.
Gruenfeld has a multitude of mainlines:
Exchange setups: (cxd5 Nxd5 e4 Nxc3 bxc3) White plays Nf3 and Be3, White plays Nf3 and Rb1, White plays Bc4
Bf4 setups
Bg5 setups
Fianchetto Gruenfeld setups (g3/Bg2)
e3 setups
The Russian System (Qb3)
and that's without going into the multitude of important sidelines like 5.Bd2 in the exchange, etc. etc.
I am looking into the Grunfeld because I am a little tired of playing KID...but goodness, this opening has so much to be learned.
Does the Grunfeld really have more to learn than the King's Indian? It seems to me that the King's Indian has a lot more to learn, and that the Grunfeld is more straightforward, but perhaps I'm just missing something. (?)
KID does have tons of learn, but since I'm playing the 1400 - 1700 crowd, we go out of book a lot earlier...so the variations are a bit less with KID, I find...esp. near the beginning. Also, I feel the Grunfeld is more punishing when one side makes a mistake...maybe just my (very limited) experience?
I see the battle is over White's d4 pawn for the most part...but could someone explain what happens if White just decides to take Black's c pawn?