How to tackle the Caro Kann Exchange: an urgent recommendation (Opening Survey)

Wouldn’t it be great to have a line in your black repertoire against which the white players fail again and again? They play what they assume are the logical moves only to realize they`re in trouble when it’s too late.

Well, check this out:


A position from the Caro Kann Exchange Variation that has been trending lately on all levels. Once White goes for the Exchange against the Caro Kann there is a huge chance you will have this on the board.


Now look at the stats:

Most white players (especially those below 2.400) are doomed to fail at this point according to the megabase.

Since White has prepared Bf4 by playing Ne2, most of your opponents will stick to this plan. They will play 7.f3 Bd7 8.Bf4 (especially since Schandorff’s “Grandmaster Repertoire: Caro Kann” book gives it as the main line). Enjoy to see your oppenents disappear into the tank once 8…e5! is on the board. Black may already be better at this point. For sure he has the easier game. The results speak for themselves:


There is a huge variety of lines and approaches for Black to fight the Exchange Variation, each of them playable. The hottest of them made headlines lately during the US championship when Sam “Captain America” Shankland fired out 8.h3 instantly against Awonder Liang’s seemingly spectacular and brandnew 7…e5!? and cruised towards an easy win from there on:


If you consult the chess author of your choice you will find many recommendations what to do against the Exchange, each of them reasonable. But from a practical standpoint (especially if most of your opponents are below 2.400 elo) there can only be one line: 5…Qc7, heading for the position in the first diagram in which White is doomed to fail.

We produced an up to date survey on 5…Qc7 in the Caro Kann Exchange that features the most important ideas and latest developments in this line. But let’s see a brief history of the whole Exchange variation first for you to make sure 5…Qc7 is the line you want to play.

A brief history of the Caro Kann Exchange

1. e4 c6 2. d4 d5 3. exd5 cxd5 4. Bd3 Nc6 5. c3 Nf6 6. Bf4


“Knights out first” ist what beginners learn, and in most cases that’s good advice. The Caro Kann exchange with its typical Karlsbad structure is different, though. With 4.Bd3 White has made sure the Bc8 can’t occupy the b1-h7 diagonal while he is occupying the h2-b8 diagonal himself with the Bf4. It can’t be a bad idea to bring out the bishops first if you can secure them the most active squares this way.

Since the b1-h7 diagonal is occupied, one major question for Black is what to do with his potentially “bad” Bc8. Back in the day they tried to exchange it against the potentially “good” Bd3 via b5 like José Raúl Capablanca did successfully in 1926 against Geza Maroczy:

7…Bg4 7. Qb3 Na5 8. Qa4+ Bd7 9. Qc2 Qb6 10. Nf3 e6 11. O-O Bb5


But what if White manages to prevent the exchange of bishops and keep the Bd7 locked in behind the black pawn chain? In 1970 against Tigran Petrosian none other than Bobby Fischer came up with a novelty that achieved exactly this.

After 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. Bd3 Nc6 5. c3 Nf6 6. Bf4 Bg4 7. Qb3 Na5 8. Qa4+ Bd7 9. Qc2 Qb6 10. Nf3 e6 Fischer introcduced the novelty  11. a4!


Yes, Black can go 11…Qb3 or 11…Nb3, but he won’t achieve anything by doing so. The point of Fischer’s move is that the Bd3 is now superior to his counterpart on d7. The game isn’t decided or anything, but White should enjoy a slight, stable edge from here on. To this day, Fischer’s novelty still stands, so that most black players try a different approach.

7… Qc8 (or 7…Dd7) have gained popularity especially after Fischer “killed” the 7…Na5 line.


Here Black intends to keep his light squared bishop outside the pawn chain. It took the white players some time to understand that they shouldn’t go h2-h3 here (even 2.400 guys still do up to this day). h2-h3 is a waste of time since Black wants to move the bishop anyway.

In a game between Vladimir Kramnik and Vassily Ivanchuk during the 2017 World Cup play continued

8. Nd2 e6 9. Ngf3 Be7 10. O-O Bh5 11. Rae1 Bg6


The crucial maneuvre for Black who manages to exchange the bishops on the diagonal b1-h7. White can argue that he maintains a slight edge nevertheless. While Black has trouble to arrange his typical Karlsbad queenside minority attack or get the break …e5 in he may face a growing attack on the kingside. Black should be ok, but he can’t afford to play carelessly.

Vladimir Kramnik handled this affair like the drunken machine gunner that he has become lately. After 12. Bxg6 hxg6 13. h4 a6 14. c4 a huge fight ensued in Kramnik,V (2803)-Ivanchuk,V (2728), Tbilisi 2017, 0-1 (71).

7… e5!? is the hot shit.


8. h3

(8. dxe5 Nh5 9. Be3 Nxe5 10. Bb5+ Nc6 was fine for Black in Wei,Y (2743)-Navara,D (2740), Yancheng 2018, 0-1 (61).)

(8. Bxe5?! Qe7)

(8. Qxb7!? has never been played, leads to an unclear mess. 8… exf4 9. Qxc6+ (9. Bb5 Rb8 10. Bxc6+ Nd7 11. Bxd7+ Bxd7 12. Qxd5 Rxb2) 9… Bd7 10. Qa6 Rb8 11. Nf3 (11. Qxa7? Bd6 and White is toast.) 11… Rxb2 12. O-O)

8… Na5 9. Qc2 exf4 10. hxg4 Nxg4 11. Qe2+ Qe7 12. Nh3 f3 13. Qxe7+ Bxe7 14. g3


has turned out to be the main line. White can argue that he will develop play on the kingside and/or pressure against d5. But it will take him some time to arrange his guys and collect the pawn on f3, time that Black can use to occupy the e file and/or advance on the queenside for instance. Black did well in the few high level games this line has seen so far.

So much for a brief history and state of affairs of the Caro Kann Exchange. Now let’s look at the highly recommended 5…Qc7 in some depth. You find all the lines below in the BodenseeBase in order to do further research and/or drill them.

Opening Survey: 5…Qc7 in the Caro Kann Exchange

1. e4 c6 2. d4 d5 3. exd5 cxd5 4. Bd3 Nc6 5. c3 Qc7


Preventing Bf4.

6. Ne2

Insisting on Bf4.

(6. h3 White prevents …Bg4 instead. 6… Nf6 7. Nf3 g6 8. O-O Bf5 9. Re1 (9. Bxf5 gxf5 10. c4 Bg7 (10… Rg8!? is an interesting, very aggressive suggestion among the machine’s top moves. Black prevents Bg5, occupies the g file, plans …0-0-0 and will transfer the Bf8 to e7 or d6 later. 11. Qe2 dxc4 12. Qxc4 e6 13. Nc3 O-O-O with a complicated position.) 11. Nc3 Rd8 with unclear play in Efimenko,Z (2708)-Riazantsev,A (2679), Olginka 2011, draw (18).) 9… Bxd3 10. Qxd3 e6 11. Nbd2 Bg7 12. Nb3 O-O and Black wasn’t worried at all in Zvjaginsev,V (2660)-Aronian,L (2784), Berlin 2015, 0-1 (68).)

(6. Na3 a6 7. Nc2 Nf6 8. h3 e6 (8… g6 is possible as well.) (8… e5 has been played by Grischuk among others, but it seems to me that the somewhat weird knight on c2 is more justified in an IQP than in a Karlsbad structure, so I’m not a fan. Grischuk may know better of course.) 9. Nf3 b5 10. O-O Bb7 11. Re1 Bd6 and things could have gone either way in Kramnik,V (2812)-Fedoseev,V (2726), Dortmund 2017, 0-1 (29).)

(6. Nd2 White passes on the option Bf4 and choses to place his knights in a natural way instead. 6… Nf6 7. Ngf3 Bg4 8. Qa4

(8. Nf1 e6 9. Ng3 Nh5 10. O-O Bd6 11. Re1 Nxg3 12. hxg3 Anand went for an equal game here with the typical …Bg4-h5-g6, but the machine suggests that Black should already be more ambitious by going …0-0-0 and …h5. Aronian,L (2764)-Anand,V (2759), Paris 2018, 1-0 (42).)

(8. O-O is the old, less ambitious move. 8… e6 9. Re1 Bd6 10. Nf1 O-O 11. Bg5  (11. Ng3 Rfe8 12. h3 Bxf3 13. Qxf3 e5 and Black is already better.) The possibility 12.Lxf6 is nothing to worry about. Most black players go for 11… Nd7 nevertheless. Black is fine anyhow. 12.Bh4 Bf4 13. Bg3 Nf6 14. Bxf4 Qxf4 15. Qd2 and draw agreed in Jussupow,A (2595)-Gustafsson,J (2615), Germany 2005.)

8… e6 9. Ne5 Bd6 10. Nxg4 Nxg4 11. h3 Nf6 12. Nf3 O-O 13. O-O Bf4 and White has managed to keep his light squared bishop unopposed at least. Not much going on besides that, though. Vitiugov,N (2721)-Vachier Lagrave,M (2811), Germany 2016, ended in a draw after 28 moves.)

6… Bg4


Preventing Bf4.

7. O-O

Has been played 207 times only according to the Megabase, but is the better move than the inferior 7.f3 that has been played 630 times.

(7. Bf4?! Qxf4)

(7. f3 White assumes that this is the logical continuation, setting up Bf4 once again. But f3 weakens the white position and disturbs the natural development of the light pieces first of all. 7… Bd7 8. Bf4 e5! is already pleasant for Black. 9. dxe5

(9. Bg3 Bd6 (9… h5!? would be a novelty) 10. Na3 a6 11. Nc2 Nge7 12. O-O h5!? and the black position looked more harmonious ins Gonzalez Acosta,B (2423)-Sasikiran,K (2668), Mallorca 2004, 0-1 (34).)


9… Nxe5 10. O-O Bd6 11. Nd4 Ne7 12. Na3 (12. Bc2?! Too passive. Black goes for it immediately: 12… f6 13. Nd2 g5 followed by long castling with a complicated game ahead, in which Black enjoys nice prospects. Brynell,S (2462)-Luther,T (2580), Plovdiv 2003, 0-1 (42).) 12… a6 13. Nac2 O-O-O!N The engine suggests to approach this aggressively and sees Black ahead already. (13… O-O has been played by grandmasters so far and is good enough for a pleasant game.) 14. Ne3 f6 (diagram) and Black is about to launch a kingside attack.)

7… Nf6

More flexible than 7… e6, but most times both of them will transpose into each other.

8. Qe1 e6

(8… e5 is possible, but less beefy in this scenario.)


9. f3

(9. h3 Bh5 10. Nf4 An idea by the rising Polish star Kacper Piorun. Black should answer it with 10… Bd6 11. Nxh5 Nxh5 and the option …Nf4 justifies to give the bishop and retake with a knight on the rim.)

9… Bh5 10. Qh4


Threatening 11.Nf4. This should be one of the key positions of the 5…Qc7 Caro Kann exchange, a tabiya even. But in more than 2.500 games it has only been reached 13 times! The main reason most white players don’t get this far (besides trying something else than 6.Ne2) is that they already run into trouble at move 7 by going 7.f3/8.Bf4.

Black has a choice now between 10…Rg8 (played by Caro Kann authority Dreev among others, and Dreev usually knows best when it comes to the Caro), 10…Bg6 and 10…h6. I’m not sure about their objective value, but to me 10…h6 with the counterthreat …g5 looks most logical. I was pleasantly surprised to see that it has been the only move played in correspondence chess, so I give it as the main line. 10…h6 has a downside, though: Sooner or later Black will have to go …Bg6 and take back on g6 with the f pawn, but it seems that is something he can afford to do.

10… h6


(10… Rg8 11. Nf4 (11. Bxh7!? may be the most critical test for 10…Rg8, gets incredibly complicated, but has never been played. This would require a ton of analysis that someone else will have to do.) 11… Bg6 12. Nxg6 hxg6 13. Na3 a6 (diagram) 14. Qe1?! A misguided attempt to keep the pair of bishops. (14. Bf4 Bd6 seems neccessary, and Black should be fine.) 14… Bd6 15. g3 Rh8 16. Kg2 and Black was clearly ahead in Alekseev,E (2617) -Arutinian,D (2536), Batumi 2018, 1-0 (66).)

(10… Bg6 11. Bxg6 fxg6 12. Re1 Kf7 With no white knight having access to e5/g5 this is much better for Black than it looks at first glance. 13. Bf4 Qb6 14. Nd2 e5 and Black was doing well in Cornette,M (2602)-Risting,E (2322), Fornebu 2017, 1-0 (38).)

11. Qh3 

Where else? You don’t want to go 10.Qh4 and then move back along the same diagonal.


(11. g4?! White refuses to back down for a move, a poor decision. 11… Bg6 12. Bxg6 fxg6 13. Nf4 Kf7 14. Nd3 Bd6 15. f4 g5 16. Qe1 (diagram) While being incredibly hard to handle for a human who has never seen this before, objectively the black position is very nice. Ne5+ is coming, but to no great effect. Black has the choice even to either walk his king to the queenside or hide it at g8/h7 for the moment. In Svoboda, S (2434)-Cernousek,L (2462), Ceske Budejovice 2018, Black hastily released the tension with 16…gxf4?! and the players agreed to a draw, probably because they had no idea who is better or worse.)

11… Bd6 12. Ng3


12… Bxg3

(12… Bg6!? 13. Bxg6 fxg6 14. Qxe6+ Kf8 is a surprising pawn sacrifice suggested by the machine, an interesting one at least. While …Re8 and …Bxg3 is in the air every single white piece sucks.)

13. Qxg3 Qxg3 14. hxg3 Bg6


15. Bxg6

(15. Be2 They do love their pair of bishops, sometimes too much. 15… h5 16. Nd2 Nd7 17. Nb3 f6 18. Kf2 Bf5 19. Be3 Kf7 20. a4 with an early draw in Santos Ruiz,M (2505)-Sorokin,A (2486), Mamaia 2017.)

15… fxg6 16. g4 g5 17. Na3


It’s a long way to e5, and he won’t be there in time.

17… Kf7 18. Nc2 Kg6 19. Be3 h5 20. gxh5+ Nxh5 21. g4 Nf4 22. Ne1 e5 23. Bxf4 exf4


The position is equal.

Draw agreed in Lovakovic,F (2348)-Tazelaar,H (2373), ICCF email, 2017.

5 1 vote
Article Rating
Benachrichtige mich bei

Most Voted
Newest Oldest
Inline Feedbacks
View all comments

[…] boys from Chessable had it exclusively in their commercial Caro Kann repertoire until it appeared for free in our survey a few days […]


[…] Liang. Die kannten wir am Bodensee zum Glück schon in- und auswendig, weil sie Teil unserer Eröffnungsübersicht zur Caro-Kann-Abtauschvariante ist. Darum haben wir vorgespult, um zu gucken, was noch kommt, und am Ende des Videos ein […]


[…] kompliziert und heikel ist, aber objektiv okay für Schwarz. Wer es genauer wissen will, guckt hier oder in der […]