In this position, White is threatening to play some combo of Bxc5 / Nxe5:

[FEN ""]
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 O-O 6.Be2 Nbd7 7.O-O e5 8.d5 Nc5 9.Qc2 a5 10. Bg5 h6 11.Be3

Here 11...b6 is the most popular response. My question is, why not Ne8 / Nfd7 / Nh5, since those moves simultaneously deal with White's threat while making way for the f-pawn to be pushed? Initially, I thought the reason was because it'd be somehow bad for Black if White decided to capture on c5 and Black had to capture back with the d-pawn. But consider this very similar line:

[FEN ""]
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 O-O 6.Be2 Nbd7 7.O-O e5 8.d5 Nc5 9.Qc2 a5 10. Be3 Ng4!

This has an ~80% winrate for Black according to the lichess Masters database, and here he's basically saying to White "please, go ahead and capture my knight! I'll capture back with the d-pawn and I'll be super happy with my position!"

So why is 11...b6 preferred to a knight move in the original position?

3 Answers 3


This is the old, old main line that is usually considered dead equal.

I can't believe the 80% winning percentage for black has to do with the objective merits of the position.

Note that Lichess' opening explorer does not show the ratings involved in those games, nor the dates. Looking at percentages only can be extremely misleading.

I think the following is happening: theory probably says that Bg5 10.Bg5 h6 11.Be3 is slightly more precise than 10.Be3, because it has provoked the weakening move ...h6, and that can sometimes be exploited.

So in the games where 10.Be3 is played, white is probably weaker on average, or he would have played 10.Bg5. That can explain the 80% score for black.

In similar fashion, the difference in frequency between ...Ng4 and ...b6 may have something to do with the time period, one was probably in fashion for a while, then the other.

I only checked Gallagher's Play the King's Indian, which contains the position after Bg5 h6 Be3 by transposition.

His assessment of ...Ng4 there is that it's obviously critical because it was played by Fischer (!) but unnecessarily risky, and he analyzes ...Nh5 and ...b6 to comfortable positions for black.

As that was a very influential book, maybe ...Ng4 became less popular then (around 2004) and ...b6 more popular?

Anyway my feeling is that it's all a matter of taste.

  • "I only checked Gallagher's Play the King's Indian, which contains the position after Bg5 h6 Be3 by transposition." -- Can you tell me what page you found it on? Thanks!
    – James Ko
    Commented Aug 5, 2021 at 3:01
  • Page 86, with the move order 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.d5 Na6 8.0-0 Nc5 9.Qc2 a5 10.Bg5 h6 11.Be3. Commented Aug 6, 2021 at 8:32

After 10. Ng4 then simply 11. Bxc5. You lost a centralized knight but more importantly, after dxc5, white can double on the d file and try to open the centre after rd1

(Back to the original position) 11....Nh5 then white can play 12. a3 (b4 in mind) taking initiative, 12. Bxc6 (rd1 in mind) threatening to open the d file, 12. Qd2 with the same idea or probably g3 denying f4 for the knight.

11.... Ne8 is passive. White probably won't play Bxc6 because otherwise the knight goes to d6. I dont think it is time to play it and f4 but since we are humans, practical chances exist. **

11... Nfd7 would take care of Bxc5 but i dont feel like it is the time to play it.

11... b6 would take care of Bxc6 and lock the position.

B6> Nfd7>Nh5>Ne8

White has a space advantage and has a slight lead in development. Opening the position would work in white's favor. That is why i think the key move in this line is 8. D5. White shouldn't have closed the position while black should try to do it. That is why b6 is preferred.

.enter image description here

White suddenly scored lower than black after 8. D5 A computer might be able to handle blacks kingside attack but a human make mistakes. Be3 is much better. If you don't open the position, attack would come from the flank. If that flank is where your king is in, it suddenly becomes more effective than an attack on the other flank due to checks and mate threats. I would think that is why black scored high

(Note that if the resulting position is the same after either Nh5, Nfd7, if Ne8 then it shouldn't matter result wise. But some are better to sidestep some ideas from white)

enter image description here

Edit: i would think statistics can describe practical chances well but there are stuffs happening too. It could be that even tho the moves are inaccurate, the opponent doesn't respond properly, the number of games are small and so on

  • Regarding your first couple of paragraphs: note that I was saying Ng4 was a strong move after 10.Be3, not after 10.Bg5 h6 11.Be3 due to the line you mentioned. After 10.Be3 Ng4 11.Bxc5 dxc5, h3 doesn't do much because the knight can retreat to h6.
    – James Ko
    Commented Jul 27, 2021 at 14:47
  • Also, what you said about Ne8 being passive / Nfd7 blocking the bishop: wouldn't that also apply to this position? Ne8 / Nd7 are the most common moves there, so why is this different? The knight will likely end up back on f6 anyhow.
    – James Ko
    Commented Jul 27, 2021 at 14:56
  • After 10.... Ng4, then simply 11 Bxc5 dxc5 and white is slightly better.
    – Rizal
    Commented Jul 27, 2021 at 15:49
  • I'm not sure if that's true; after 10...Ng4 the Lichess Masters database records 80% wins for Black.
    – James Ko
    Commented Jul 27, 2021 at 16:17
  • Ya because we are humans. 1 Na3 scored high but obviously most won't play it. Due to the kingside attack, it is no wonder why black scored high. That is why white shouldn't have closed the position with d5 in the first place. You can see black winning more after 8. D5 as opposed to 8 Be3.
    – Rizal
    Commented Jul 27, 2021 at 21:57

Rizal provides a good move analysis. Positionally, the black knight at c5 is well placed and moving it would decrease its influence. Supporting the knight with b6 keeps the knight on a good square. If white takes the knight, then capturing the bishop with bxc5 keeps the center closed and opens the b file for black. This makes white's typical plan of queen side expansion difficult. Black can proceed with the usual f7-f5 plan unhindered by white's queenside pressure.

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.