Here is a recent game in the King's Indian I had as Black:

[Event "Rated Rapid game"]
[Site "https://lichess.org/za0zcaTi"]
[Date "2021.08.23"]
[White "tomx2"]
[Black "fian-ketto"]
[Result "1-0"]
[UTCDate "2021.08.23"]
[UTCTime "15:47:45"]
[WhiteElo "1945"]
[BlackElo "1967"]
[WhiteRatingDiff "+8"]
[BlackRatingDiff "-7"]
[Variant "Standard"]
[TimeControl "900+10"]
[ECO "E71"]
[Opening "King's Indian Defense: Makogonov Variation"]
[Termination "Normal"]
[fen ""]

1. d4 g6 2. e4 Bg7 3. c4 d6 4. h3 Nf6 5. Nc3 O-O 6. Nf3 e5 7. d5 Nh5 8. Bg5 Qe8 9. Bd3 h6 10. Be3 Nf4 11. O-O Na6 12. Bc2 f5 13. a3 Nc5 14. exf5 gxf5 15. b4 Nd7 16. Bxf4 exf4 17. Re1 Qh5 18. Qd2 c6 19. dxc6 bxc6 20. Re7 Nb6 21. c5 Nc4 22. Bb3 d5 23. Bxc4 dxc4 24. Rae1 Rf6 25. Re8+ Kh7 26. R8e7 1-0

After move 10, I had the option of taking the light-squared bishop with my knight. However, considering it was one of my only developed pieces and I had just moved it 3x out of the opening, I decided to keep the tension and develop another piece instead. However, according to Stockfish my 11th move is a terrible blunder; the evaluation swung from slightly negative to +1.1, and stayed positive for the rest of the game.

Why is it desirable to trade my very active knight for his bad bishop that's blocked in by all of his pawns?

4 Answers 4


This is an interesting question. You were basing your play on some useful truisms without considering if they really applied. "I need to free my position with f5." "I should make Pawn captures toward the center". "He will not want to give up his good Bishop", "I should not waste tempi" All true, individually, but almost nothing is true every time, which is why chess rewards creativity. Before basing your play on general principles, it is wise to ask why do those principles usually apply? and do they apply in this position? GMs giving simuls rely on general principles because they must, but they do not do so thoughtlessly.

The key here is the resulting pawn structure. Botwinnik famously asserted that "every Russian schoolboy" knows to recapture on f5 with the Pawn, but that was in a position where Black ends up with Pawns on e5,f5, that are mobile and control the center. In this game White was able to play Bxf4, when the doubled pawns on f5,f4 are static, vulnerable, obstructive and impotent.

At move 10 note that Bxf4 is not yet a threat because it gives up the good Bishop, creates a strong square for you on e5 and sets you up with a K-side Pawn-roller. It also fundamentally changes the position so that your plan may no longer revolve around ..f5, but around ..g5 instead. Why would you free up the Bd3?. Always remember that plans can change depending on your opponents play. In the game, White delayed Bxf4 until after you had played ..f5 and had recaptured with the Pawn, creating that awful self-inflicted Pawn structure.

So at move 10 it was OK for you leave the K-side alone and look for play elsewhere, which you did. Unfortunately ..Na6 does not fit with any long-term plan because you can be pushed from c5 by b4 and end up taking three moves to get to d7. The plan of N-a6-c5 is workable, if you play ,..a5 first.

But also it would have been good to exchange on d3 despite the loss of tempo. This is a closed position and will remain so for some time. Tempi only matter if they can be used to create threats. As Black after 10 moves you have as nice a position as you can expect. Whites play has been sound but unambitious and you can take your time. There is nothing complex here that needs deep analysis. Objectively the game is equal, but psychologically it favors Black because has attack is on the King-side and "mate ends the game"

Which is not a reason to hurry anything. Rather than trying to use tactics to improve the position it is often better to improve the position as much as you can before embarking on tactics. The question is then, are there still ways for me to improve the position slowly? or do I grab at the tactics whilw they are there? A slow approach to the position at move 10 is to play ..a5, Na6, Nc5 and only then ..f5 when Blacks forces are focussing on e4 and forcing some kind of resolution in the center. After this Black will continue preparing his attack with moves like Q-h5 or g6, and Rae8. Attempts at "analysis" are meaningless at this stage. All you need to know now is that there will be plenty to do later on.


Done a little bit of analysis on the position using a stronger engine (Stockfish 14) and honestly I wouldn't put much weight in it. I wouldn't be stressing this at all. Unless you really stress about the engine analysis at a depth of 27 with a difference of -0.17, basically 1/5th of a pawn.

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[SetUp "1"]
[FEN "rnb1qrk1/ppp2pb1/3p2pp/3Pp1Bn/2P1P3/2NB1N1P/PP3PP1/R2QK2R w KQ - 0 1"]
[PlyCount "6"]

1. Be3 Nf4 2. O-O Nxd3 3. Qxd3 f5 *

At the f5 position this is how I review the board.

  1. Queen is Overloaded defending e4 and c4.
  2. f5 threatens fxe4 , if exf5 bxf5 threatening the queen. White should respond by pinning its own knight to the queen or retreating. Either way , you develop with tempo and white is forced to repeatedly move the same pieces around.

I agree with the Nxd3 recommendation but i dont think your move was "blunder" by any means.

Engines are a bit of a catch 22. They can greatly assist analysis but also just confuse things. We aren't computers we can't analyse things at 4k+ kN/s and see that board at a 30+ move depth. When they are reviewing that position they are looking at 20 to 30 moves down the road to say yep...black is winning by 1/5th of a pawn.

  • At a depth of 33 the position is still equal @ -0.11 , {[#]} 1. Be3 Nf4 2. O-O Nxd3 3. Qxd3 f5 {Stockfish 14:} 4. b4 Nd7 5. a4 Nf6 6. Nd2 a5 7. Nb5 Qd8 8. f3 axb4 9. c5 dxc5 10. Qb3 Rf7 11. Bxc5 Nh5 12. Rf2 b6 13. Bxb4 Nf4 14. Kh1 Ba6 15. Raf1 Qh4 16. Nc4 Bxb5 17. axb5 fxe4 18. fxe4 Ra1 19. Qf3 Ra4 20. Rb2 Rf6 21. Be1 Qh5 22. Qb3 {-0.11/33} *
    – Dheebs
    Commented Aug 24, 2021 at 3:49

I'm not a highly rated player but I believe that after whites pawn push (around the 20th move) the diagonal to the king opens up and there isn't even a pawn to cover it up so I guess that is a weakness for blacks and they need a piece to control the diagonal along with the queen because in diagonal attack it would be a key component. Stockfish was just analyzing future moves and it understood that the bishop on the diagonal would be powerful in the attack because the bishop may seem like a bad one but it can become a powerful one as seen later in the game


Thanks to @Philip Roe for the wonderful and detailed answer. I took a look at the 9...Nh5 variation of the Bayonet Attack, which contains some lines similar to this that I found insightful and basically confirms what you say.

[FEN ""]
1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. Nf3 O-O 6. Be2 e5 7. O-O Nc6 8. d5 Ne7 9. b4 Nh5 10. Re1 Nf4? 11. Bf1 f5 12. Bxf4 exf4 13. e5

The main line is 10...f5, with a high winrate for Black after White chooses 10...Nf4. As far as I can tell, this is because 10...Nf4 is a waste of time; with 10.Re1 White has indicated that they're simply going to meet it with 11.Bf1. Also, it actually gets in the way of our plan of f5-f4 for multiple reasons: (1) as soon as we commit to f5, White can play Bxf4 to inflict doubled f-pawns on us, which basically renders our previous move useless since the f5-pawn isn't going anywhere, and (2) even if White chooses not to capture, we can't actually play f4 because our knight is in the way!

Another illustrative line:

[FEN ""]
1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. Nf3 O-O 6. Be2 e5 7. O-O Nc6 8. d5 Ne7 9. b4 Nh5 10. c5 Nf4 11. Bxf4 exf4 12. Rc1 h6 13. Nd4 g5

This seems to confirm your claim that if White decides to snap off the knight with Bxf4, then due to the change in pawn structure the correct plan switches from one centered around f5 to one centered around g5.

Here are my key takeaways:

  • In the main line Bayonet Attack, the reason you play 9...Nh5 is to (make way for your f-pawn, obviously, and) threaten 10...Nf4 to win the bishop pair. If White wants to save their two bishops, then they have to waste a move with 10.Re1 or 10.g3, giving us time to carry out our plan with 10...f5.

  • We don't actually want to play 10...Nf4 unless we know that it will force White to give up his bishop pair (ie. in the 10.c5 line), because otherwise the knight just gets in the way of our usual f5-f4-g5 plan.

  • When White decides to play Bxf4, the correct plan switches from one centered around the f5 thrust to one centered around the g5 thrust.

  • Thank you for the compliment. Im glad my remarks made sense. The Kings Indian often involves difficult judgement calls. If you get them right you will score well.
    – Philip Roe
    Commented Aug 26, 2021 at 22:33

Your Answer

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

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