In round 3, Firouzja and Artemiev played a very interesting Caro-Kann game, with white constantly trying to open up the position even at the cost of pawns, in order to exploit black's exposed king position, which indeed proved to be decisive and white won.

It seems that in these Caro-Kann structures, white has to be really careful not to lose the center pawns with some combinations. For example in the game, after white castled, I noticed that although at first glance it seems the e5 pawn can be won after Nxg3, fxg3, Nxe5 using the pin, but white has the Na4 intermezzo first dislocating the queen then winning the knight on e5. Here it is diagrammed:

 [title "e5 not really hanging"]
 [fen "r3kb1r/pp1n1pp1/1qp1p1b1/4PnBp/2BP3P/2N3N1/PP3PP1/2RQ1RK1 b kq - 2 12"]
 [startflipped "0"]

 1...Nxg3 2. fxg3 Nxe5 3. Na4 Qc7 4. dxe5

This one I could figure out. But what I don't get is, instead of going after the e5 pawn, after castling, couldn't black have won the d4 pawn which is doubled attacked by the knight on f5 and the queen on b6, while only being defended once with the queen:

 [title "why didn't black take on d4 here?"]
 [fen "r3kb1r/pp1n1pp1/1qp1p1b1/4PnBp/2BP3P/2N3N1/PP3PP1/2RQ1RK1 b kq - 2 12"]
 [startflipped "0"]

  • Why couldn't black take on d4? Is it simply because after Qxd4 Qxd4 Nxd4 white has sufficient positional compensation considering all the pieces are actively posted and black's still behind in development? or are there simply direct tactics preventing Qxd4 altogether?

1 Answer 1


Good job on the first tactic: indeed e5 is not really hanging in that line due to Na4.

As for the d4 pawn, although even in the line you suggest (with the exchange of queens) white stands a tad better, the d4 pawn is in fact tactically defended! Considering that you do not mention any computer evaluations and have tried to tackle the problem on your own (which is excellent!), I reckon you're more interested in not just seeing the solution but to also see how to get there, so let's walk through it together.

To recognise the combination, take note of black's weakpoints around the king:

  • king stuck in the center,
  • the dark the square holes around black's king,
  • the bishop taking away the e7-d8 squares, and the potential of a smothered type of mate were the knight to land on c7,
  • the open d-file (thanks to Qxd4) and the fact that the d7 knight is only defended once by the king (therefore vulnerable to direct tactics).

With these in mind, now try to find forcing moves (captures, checks, threats of mate or gain of material, etc.) that exploit these weaknesses and see if you can make them work concretely.

The first move that comes to mind, challenging all the aforementioned weaknesses at once, is Nb5, with the double threat of attacking the queen and mating on c7. Rather interestingly, Nb5 works out beautifully: if black doesn't capture on b5, then they lose on the spot due to dark square weaknesses. And if they do capture the knight, then they end up further opening the position (as cxb5-Bxb5 open the critical a4-e8 diagonal towards the king) and lose due to the compromised king position and lack of defenders (in turn caused by lack of development).

Let's see how all this pans out concretely: (no time to cross check the lines with a computer, I leave that up to you!)

 [title "Lines after 12...Qxd4"]
 [fen "r3kb1r/pp1n1pp1/1qp1p1b1/4PnBp/2BP3P/2N3N1/PP3PP1/2RQ1RK1 b kq - 2 12"]
 [startflipped "0"]

 1...Qxd4 2. Nb5 cxb5 (2...Qxd1 3. Nc7#) (2...Qb6 3. Nxf5 exf5 {loses on the spot as it allows e6 opening the e-file} (3...Bxf5 4. Nd6+ Bxd6 5. Bxe6 {A real blow to black's position! Note Qxd6 at once doesn't work due to Qc5} fxe6 (5...Bxe6 6. Qxd6 {now Qxd6 works as the c5 square is guarded by the rook on c1}) 6. Qxd6 {only way to avoid mate without losing the queen is to make a run for it via f7} Kf7 7. Qxd7+ Kg8 {and white's clearly winning thanks to the domination of the bishop and queen, e.g. after Rcd1 Kh7 Qf7 wins the e6 pawn}) 4. e6 Nf6 5. exf7+ Bxf7 6. Re1+ Ne4 7. Nd6+ Bxd6 8. Qxd6 {and black has to sacrifice their queen to avoid mate}) 3. Bxb5 a6 (3...Qxd1 4. Rfxd1 f6 5. exf6 gxf6 6. Rxd7 a6 {threatening to win the rook after Ba4 b5} 7. Ne4 {Nf6 threatened with mate} Be7 8. Rxe7+ Kxe7 9. Bxf6+ Kf7 10. Rc7+ Kf8 11. Bxh8 axb5 12. a3 {and white's completely winning, being up a healthy pawn, better piece coordination, while black's king remains exposed and the 3 isolated pawn islands are useless offering no prospect of counter-play.}) 4. Qxd4 Nxd4 5. Bxd7+ Kxd7 6. Rfd1 f6 7. Rxd4+ Ke8 8. exf6 gxf6 9. Bxf6 Rh7 10. Re1 Rc7 {baiting white to take on e6 with Rxe6+ Kf7 and white's rook defending the bishop will be easily overloaded} 11. Bg5 {calmly side-stepping the double attack with Kf7} Bf7 12. Ne4 {and white's advantage is once again decisive: clear pawn up, all 4 pieces perfectly coordinated against black's king stuck in the center, meanwhile all of black pieces are bound defensively, and both the e6 and h5 pawns are extremely weak and will fall soon with slow technical play.}

What a beautiful tactical adventure! This makes for excellent practice of both tactics and calculations.

It's worth noting that in some of the lines discussed above, black is not necessarily getting checkmated or even being necessarily a piece down, instead, the position becomes heavily damaged and black ends up at least a clear pawn down, all of which is more than enough of an edge to win considering the caliber of players we are talking about.

Finally, Artemiev need not necessarily have calculated all the above lines to realise d4 is not in fact hanging. Instead and particularly for players of such strength, they either have seen it before in a game or in preparation, or based on their strong intuition they simply smell the overwhelming dangers that could arise if they went pawn hunting with their king stuck in the center.

Last but not least, calculating very forcing lines (such as the above) is in fact the easiest type of calculations due to their inherent nature of being forcing (their branch outs are limited and the remaining lines are very concrete thus requiring less positional judgement). Meaning that e.g., performing all the above calculations over the board is quite feasible for strong players, and somewhat even trivial for super GMs.

  • Another nice answer. Just a note: This was all book up to 11...Nf5, and in fact, the alternative 11...Nd5 was played in the recent Rapid World Championship. It being so early in the game, I strongly suspect that Firouzja has had this on a board in prep, and even if Artemiev has not, he just "trusted" his incredible opponent enough not to take that pawn. I still suspect that he instantly saw Nb5 and that f7/d7 were so weak that he instinctively knew he could not take. Jan 14, 2020 at 16:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.