Kamsky has an almost even record against Anand but a terrible record with Topalov. If he can get past Topalov he has chances with Anand I believe but would be the underdog.
Facing the Caro Kann

Your obsession with openings will retard your progress. Against players rated under 1800 any opening will do. Learn to play chess first then when you hit 1800-1900 start perfecting you openings of choice.
And against players over 1800? I do play guys rated 1800-2000 USCF regularly, including the guy who always plays the CK who I thought I might end up facing one Saturday, who is rated over 1900. As it turns out, I faced an opponent in the 1500's and got an easy win on the white side of the Hungarian Defense. That boosts my rating up to the 1600's for the first time. Note to people who play the Hungarian as black: Don't try to transpose to a Scotch later. (1. e4 e5 2. Nf3 Nc6 3. Bc4 Be7 4. d4 exd4 5. Nxd4 Bc5?? 6. Bxf7+! and white has an extra pawn with black's king exposed)
There's another guy in the 1500's who I face regularly who also recently started playing the CK, and I'm not as worried about facing him. But I'd like to be as prepared as reasonably possible to play higher rated guys in the CK. I realize that I'm not going to beat anyone on book knowledge, but I want to know enough to reach a reasonable position and have some sort of plan against opponents who know what they're doing. I'll pick up more book knowledge later, as I play against the opening more often.
But as I tried to emphasize in my first post in this thread, I'm really looking for plans more than anything else. I'm not interested in memorizing specific, sharp opening variations. I'm happy enough playing quieter lines, as long as I know what general middle game plan I should be following once I get there. To me, opening study is all about ideas, not specific moves. For that matter, the same could be said of chess as a whole.
Well, the basic idea in Caro-Kann is to attack the e4 square, c6 is supporting d5 which is attacking e4. This idea is similar to French where it is the e6 pawn that is supporting d5. The idea with c6 instead of e6 is to give space for the bishop on c8. Well, that is just the start of course.
So we have the start: 1.e4 c6 2.d4 d5
Now, here are some choises:
Do you want to focus on e4 and defend it you can play: 3.Nc3 (or you could play Nd2) and get into a main line that you probably know.
If you want to have an open e-file against a black open c-file you can play: 3.exd cxd and thus have a game with focus on that.
Or if you want the advance variation which closes the center you can play: 3.e5 and maybe plan for an attack with the pawns g4, h4 and sooner or later f4. It can look impressive with all those pawns, but it can be rather difficult to break through black's position. It usually gives room for a lot of tactics in my experience.

Personally against the CK i would prefer the advance. the following line should give you an extreme advantage in space and cramp black.
e4 c6 d4 d5 e5 Bf5 Nc3 e6 g4 Bg6 h4 h5 g5
Pretty clear that the f6 square is fixed good and proper. f4 to reinforce your pawns is a good idea, thoguh sometimes bf4 serves as well.
Black may have a solid position, but it is also extremely cramped. Castle long, bring out your pieces and ram him with your pawns.

Gotta say, I disagree wholeheartedly with the above analysis.
after these moves
1.e4 c6 2.d4 d5 3.e5 Bf5 4.Nc3 e6 5.g4 Bg6 6.h4 h5 7.g5
I contend that black is better. Here's why:
1. There's no pawn breakthrough on the kingside for white.
2. Black has free availability of the f5 square, for the bishop and/or knight.
3. With the kingside closed off to immediate tactical threats, black can develop leisurely. While black has less space, there are enough squares for him to develop.
4. The white dark-squared bishop is just an awful piece.
I would say the position is very difficult to play for white, black has a logical plan, and I would characterize it as =/+

According to chess assistant after 7 g5 white scores less than 30% which is terrible for white. 6 Nge2 is normally played and it scores about 56% for white.

I'm guessing that your percentages take into account the draws?
Good to know that my "gut feel" analysis is borne out by the almighty silicon!

You can play the classical variation, i.e. 1.e4 c6 2.d4 d5 3. Nc3 dxe4 4.Nxe4 Bf5 5.Ng3 Bg6 6.Bc4 Nd7 7.N1e2 e6 8.0-0 ... and then will move the f pawn. The idea is f4-f5

I'd definitely agree with the verdict on the 6. h4 h5 7. g5 line. I think most of the time White's trying to concoct a pawn storm prematurely; it's all too easy for Black to block it physically with the bishop on g6, and a knight on f5 (via e7). I'd be very comfortable as Black in that position.

I played the Caro-Kann for years and years until I realized that there were holes in just about every line of my repertoire. The lines I had the most trouble with were the Two Knight's (1.e4 c6 2.Nf3 d5 3.Nc3), the 2.d3 d5 3.Nd2 variation and the Panov-Botvinnik Attack.
Incidentally, when I play 1.e4 I'm usually hoping that my opponent will play 1...a6 or 1..b6, but unfortunately, they all seem to play the Sicilian for some reason.

In the two knights variation there are some good traps in result of black bringing out his light squared bishop, and result in good solid positions. I remember watching a video from chessvideos.tv but im terribly sorry I cant remember where I saw it.

Who fears the caro kann ?
Well, the book on Bobby Fischer's 60 memorable games notes that in the 7 games in which he faced the ck he scored a slight minus score.

Caro-Kann is the only e4 defense I play nowadays.
1. I don't agree that the Panov really takes black out of "his comfort zone". Panov lines are not that threatening compared to the mainline (capablanca variation) and the aggressive advance lines. If black knows where the pieces go he will disarm white and white will be left with a weak pawn structure to an endgame. But if you really really only want to play wide open games, go ahead play it.
2. If you want to study theory and really put the CK player to the test you should look at the main line or the advance shirov variation with Nc3 followed by g4. It's a minefield. For both sides, I might add.
3. If you don't want to study theory but want an ok position with possible kingside attack, play the exchange variation.
4. Black-Dimar gambit is crap. Any CK player worth his salt will know how to develop safely without giving you a chance to exploit any tactics on f7 and grind you down with that extra pawn.
5. Fantasy variation is also weak. He will exchange and strike with e5 and develop and avoid tactics on f7 and abuse your weak dark-square diagonal towards you king.

Gotta say, I disagree wholeheartedly with the above analysis.
after these moves
1.e4 c6 2.d4 d5 3.e5 Bf5 4.Nc3 e6 5.g4 Bg6 6.h4 h5 7.g5
I contend that black is better. Here's why:
1. There's no pawn breakthrough on the kingside for white.
2. Black has free availability of the f5 square, for the bishop and/or knight.
3. With the kingside closed off to immediate tactical threats, black can develop leisurely. While black has less space, there are enough squares for him to develop.
4. The white dark-squared bishop is just an awful piece.
I would say the position is very difficult to play for white, black has a logical plan, and I would characterize it as =/+
+1
7. g5 is nothing for white. It's definately something like =/+.

Something tells me that black is instantly equal if white retreats in response to Nf6:
1.e4 c6 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nf6 5.Ng3?!
The whole point of the 4... Nd7 line is to "prepare" Ngf6, so black won't have doubled pawns if white captures. And so naturally the only critical line is 5.Nxf6+ in response to 4... Nf6

...So I've decided to try the main line - 1. e4 c6 2. d4 d5 3. Nc3 dxe4 4. Nxe4, and move that knight to g3 if it's attacked (Bf5 or Nf6). I'm sure there are sharper lines, but I don't want to focus too much on opening study.
You might also consider 5. Nc5 in response to Bf5. This provokes black into playing b6 more often than not, and from there you can play Nb3 and then a kingside fianchetto to have long-term positional attacks on black's weakened queenside. Its a simple system to remember, though as usual, the stronger player will prevail.
The thing to remember is you're not going to somehow solve the CK enough to start beating players stronger than you. You may reach a middlegame with your share of chances, but in the end, skill will give them more odds to win.
Prepare to face this line if you go 5. Nc5

Right, which has scored poorly for black and is VERY rarely played at IM/GM level. You need games you can emulate, and the b6 line has plenty of material for both sides.
The score is based on one game and there is an improvement on it in Lars Schandorffs GM Rep Caro-Kann. I doubt he gives a line that is worse then the b6 line. This line puts white to the test unlike the tame b6 one. Yes it's rarely played but I wasn't the one suggesting Nc5.

I find it amusing that you people have revived a thread from a year and a half ago.
FYI, as the guy who started the thread, I just want you to know that I don't even play 1. e4 right now. I've been playing 1. f4 for a couple of months now, to get used to it, though I plan to switch back to 1. e4 some of the time. I have Aagaard's book on the Panov-Botvinnik Attack, so I'll probably wait until I have some time to go through that, and study something better against the Sicilian than the Smith-Morra Gambit I used to play, before I switch back to 1. e4 full time.
While I would also enjoy an Anand/Topalov match, I love the way Kamsky plays - I view him as an underdog in the vein of Korchnoi.