MNb wrote on 02/03/14 at 22:18:08:


Actually, it is not irrelevant. The plan of opening the queen side is correct. However, it has been discovered that this is considerably strengthened if white first transfers his remaining knight to f3. Black will usually feel the need to play his pawn to h4. Putting the knight on f3 means that black is handicapped  in his ability to respond on the queen side as moving the bishop from e7 might mean that h4 simply falls off. Obviously, if the knight is going to f3 then it does make a big difference whether it starts from d2 or c3. You can find all this explained in Greet's "Beating Unusual Chess Defences".

tony37 wrote on 02/03/14 at 12:20:22:


That's because white's best chances in the Gurgenidze involve playing f4 and repositioning his queen's knight to f3. Obviously, if he can do this via d2 rather than c3 he will save several tempi. Black really shouldn't allow this.

This is wandering a little off topic for this thread but it is useful to look at this from a Modern move order. Assuming black wants to play a Gurgenidze structure, after 1.e4 g6 2.d4 he would really like to wait for white to commit his queen knight.  Ideally, he would prefer a move other than 2...d6 or 2...Bg7, both of which represent tempi lost in the Gurgenidze. However, after the only likely candidate 2...c6, white can also mark time by playing 3.f4.

If black now plays, for example, 3...d6 then white can continue 4.Nc3 satisfied that if black now tries 4...d5 then 5.e5 reaches a line that could be reached from the Caro move order where white has been gifted the free move f4. On the other hand, if black goes ahead and plays 3...d5 anyway, then take a look at Kosintseva-Blatny, Moscow 2004 to see how white made very effective use of the fact that she had held back her queen knight.

ErictheRed wrote on 02/02/14 at 23:52:33:


Right. There is a school of thought that 3.Nc3 is inaccurate in the Caro Kann precisely because black can transpose to a good Gurgenidze so probably Seirawan fully intended to play a Caro against 3.Nd2 and only transposed to the modern to take advantage of Smyslov's move order. So, I think your original point fully stands - this was a Caro Kann played in must win situation.

I can see some attraction of that ...Nh6 Gurgenidze for winning purposes, but of course it isn't exactly your standard Caro ...

ErictheRed wrote on 02/02/14 at 14:14:23:


In the 1985 Candidates, Smyslov and Seirawan played only one game. Smyslov was white and it was a Bogo-Indian.
http://www.chessgames.com/perl/chessgame?gid=1126817

In fact, I can't find any Caro game between Smyslov and Seirawan.

Maybee a bit off topical, but I read somewhere that when Anand needed a win as black and played against  Shirov he always plays Caro Kann

GM Shankland said something in one of his video lectures about how if you have to play for a win, and your opponent just needs a draw to get the result he wants, your best bet was to play the carokann. He then proceeded to show how he did just that in the final round of a tournament and won his game.

Aziridine wrote on 01/16/14 at 16:34:11:

Collins' quote in the same book would probably not suit your purposes:
"To be honest, I don't really see the the appeal of the Caro-Kann for non-professionals- although its main aim is to have a quiet life, White has the choice of several razor sharp lines all of which demand superb preparation. In lower-level tournaments where neither a draw with Black nor hours of preparation are always desired, the Caro-Kann just looks like a silly choice."