How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Karol Majewski
Posts: 94
Joined: Sat Jul 09, 2011 2:18 pm

How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by Karol Majewski » Sat May 04, 2019 6:04 pm

So here's the position:



It seems that this tiny 6.h3 move is the critical line of Najdorf. No wonder that it is very popular among top GMs these days. So I tried to equalize as black - of course with help of Stockfish - and couldn't find equality. I've looked at:

6...e5 7.Nde2 b5 - but here 8.Ng3 gives White slight edge.
6...e5 7.Nde2 h5 - here after 8.g3 White is somewhat better.
6...e6 7.g4 - this is mess, but I couldn't find anything close to 0.00.
6...b5 - according to my analysis, this is so far the most promising move, but I'm not sure how to assess position after 7.Nd5 e6 8.Nxf6+ Qxf6 9.g3.

So, I belive that in order to solve this, great depth and huge computing power are required. Can someone perform such analysis with Stockfish from the given position?

zullil
Posts: 5609
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by zullil » Sat May 04, 2019 10:52 pm

Karol Majewski wrote:
Sat May 04, 2019 6:04 pm
So here's the position:



It seems that this tiny 6.h3 move is the critical line of Najdorf. No wonder that it is very popular among top GMs these days. So I tried to equalize as black - of course with help of Stockfish - and couldn't find equality. I've looked at:

6...e5 7.Nde2 b5 - but here 8.Ng3 gives White slight edge.
6...e5 7.Nde2 h5 - here after 8.g3 White is somewhat better.
6...e6 7.g4 - this is mess, but I couldn't find anything close to 0.00.
6...b5 - according to my analysis, this is so far the most promising move, but I'm not sure how to assess position after 7.Nd5 e6 8.Nxf6+ Qxf6 9.g3.

So, I belive that in order to solve this, great depth and huge computing power are required. Can someone perform such analysis with Stockfish from the given position?
Stockfish-dev with MultiPV = 3:

+0.23 6... h6 7. Bd3 Nbd7 8. O-O e6 9. Qe2 Be7 10. a4 Nc5 11. Be3 Nxd3 12. cxd3 Bd7 13. f4 O-O 14. a5 Rc8 15. Qf2 Ne8 16. Nf3 Nc7 17. Rfe1 Qe8 18. Qg3 Nb5 19. Rac1 Kh7 20. Bb6 Nxc3 21. bxc3 Bd8 22. Qf2 Bxb6 23. Qxb6 Bc6 24. f5 Qd7 25. Nd4 e5 26. Nf3 f6 27. c4 Rfd8 28. Nd2 Qe7 29. g3 Qd7 (depth 46, 4:25:22)

+0.26 6... b5 7. Nd5 e6 8. Nxf6+ Qxf6 9. a4 b4 10. Be3 Bb7 11. Qd2 Bxe4 12. Qxb4 Qe7 13. f3 Bb7 14. Bd3 Qd7 15. Qd2 Be7 16. O-O O-O 17. b4 Nc6 18. Nxc6 Bxc6 19. a5 Bf6 20. Rab1 d5 21. c4 dxc4 22. Bxc4 Qxd2 23. Bxd2 Rfd8 24. Be3 Bd5 25. Bxd5 Rxd5 26. Rfd1 Bc3 27. Bc5 Rad8 28. Rxd5 Rxd5 29. Kf2 f5 30. Rc1 Be5 31. Ke2 Kf7 32. Rd1 Rxd1 33. Kxd1 f4 (depth 46, 4:25:22)

+0.27 6... e6 7. g4 h6 8. Bg2 Be7 9. Be3 Nc6 10. Nxc6 bxc6 11. Qe2 Rb8 12. b3 d5 13. Bf4 Rb7 14. O-O O-O 15. Rad1 Nd7 16. Na4 Nc5 17. Nxc5 Bxc5 18. exd5 cxd5 19. c4 Rd7 20. cxd5 exd5 21. Qd3 Bb7 22. Rfe1 Qb6 23. Re2 Rdd8 24. Bg3 a5 25. Qf3 Bd6 26. Bxd6 Rxd6 27. Qf4 Qd8 28. Re5 Qd7 29. Qd4 Re6 30. Rde1 Rc8 31. Rxe6 fxe6 32. Qb6 Rc2 33. Qxa5 Qc8 (depth 46, 4:25:22)

Nay Lin Tun
Posts: 514
Joined: Mon Jan 16, 2012 5:34 am

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by Nay Lin Tun » Sat May 04, 2019 11:25 pm

6. ...e6 7. g4 is not a mess. It is just a theory line. When white play g5, Nd7.( And when white play f4 then f5, put that kt back on Ne5 outpost).
Meanwhile keep your king in the centre, and work on your queen side.

I just did a quick search with Lco and amazed that Leela chose that theory line!!
cap4.png
cap4.png (57.47 KiB) Viewed 1214 times

todd
Posts: 44
Joined: Thu Apr 19, 2018 7:09 pm

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by todd » Sun May 05, 2019 4:12 am

I've done a lot of SF depth 50+ analyses from a variety of positions in the 6. h3 line over the last few years.

If your goal is simply to get SF (without contempt) to say 0.00 rather than having an easy position to play, then we can start from the following position (unless you think white should do something else)

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e5 7. Nde2 h5 8. g3 Be6 9. Be2 b5 10. O-O Nbd7

A) 11. a4 b4 12. Nd5 Nxd5 13. exd5 Bf5
B) 11. Nd5 Nxd5 12. exd5 Bf5

And Be7 against most other reasonable white moves, like Bg5, Be3, Re1, b3.

I'll avoid going into further detail for now because I'm not sure if your goal is "high depth SF 0.00," an Lc0 eval, or something more human.

It takes some patience to find good moves for black with SF's help in this line - the evals tend to be pretty favorable for white at low to medium depth but generally come down as depths increase.

zullil
Posts: 5609
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by zullil » Sun May 05, 2019 10:51 am

todd wrote:
Sun May 05, 2019 4:12 am
I've done a lot of SF depth 50+ analyses from a variety of positions in the 6. h3 line over the last few years.
...

It takes some patience to find good moves for black with SF's help in this line - the evals tend to be pretty favorable for white at low to medium depth but generally come down as depths increase.
Indeed:

+0.08 [+] [*] 6... e6 7. g4 h6 8. Bg2 Be7 9. Qd3 Nc6 10. Nxc6 bxc6 11. e5 Nd5 12. exd6 Bxd6 13. Ne4 Bc7 14. c4 Nf4 15. Qxd8+ Kxd8 16. Kf1 f5 17. gxf5 exf5 18. Nc5 Nxg2 19. Kxg2 Ke8 20. Re1+ Kf7 21. Bd2 Rd8 22. Bc3 g5 23. Re2 Re8 24. Rxe8 Kxe8 25. Re1+ Kf7 26. Be5 Bd8 27. Rd1 Be7 28. Nd3 Be6 29. Bc3 Bd6 30. b3 Re8 31. Nb4 Bxb4 32. Bxb4 f4 33. Re1 Bf5 34. Rxe8 Kxe8 35. Bc3 c5 (depth 51, 15:52:51)

+0.11 [+] [*] 6... e5 7. Nde2 h5 8. g3 Be7 9. Bg2 b5 10. Nd5 Nxd5 11. Qxd5 Ra7 12. Be3 Rb7 13. O-O Nd7 14. a3 O-O 15. Qd1 Nb6 16. b3 Nd7 17. Nc1 Nf6 18. Nd3 Bd7 19. Kh2 Qc7 20. f4 Bc6 21. Rc1 Bxe4 22. fxe5 dxe5 23. Rxf6 Bxg2 24. Rxa6 Be4 25. Qxh5 Bg6 26. Qxe5 Bxd3 27. Qxc7 Rxc7 28. cxd3 Rxc1 29. Bxc1 Rc8 30. Bf4 Rc2+ 31. Kh1 Rb2 32. a4 Rxb3 33. axb5 Rxb5 34. Rc6 Rb2 35. Rc8+ Kh7 (depth 51, 15:52:51)

+0.28 [+] [*] 6... h6 7. Bd3 e5 8. Nde2 Be7 9. O-O Be6 10. f4 exf4 11. Nxf4 Nbd7 12. b3 O-O 13. Kh1 Nc5 14. Bb2 Nfd7 15. Ncd5 Ne5 16. Qe2 Bh4 17. Rad1 b5 18. Bd4 Rc8 19. a4 Ncxd3 20. cxd3 Bg5 21. axb5 axb5 22. b4 Bxf4 23. Nxf4 Qg5 24. Be3 Bb3 25. Rd2 Qh4 26. Rb2 Be6 27. Rbb1 Ra8 28. Ra1 Rxa1 29. Rxa1 Bg4 30. Qf2 Qxf2 31. Bxf2 Be6 32. d4 Nc4 33. Ra7 Rc8 34. Nxe6 fxe6 35. Ra6 (depth 51, 15:52:51)

This is simply an ongoing search from the root position, with MultiPV = 3 and Analysis Contempt = Off. Perhaps 0.00 will show eventually, but, as you point out, guided searching would be much more sensible.

zullil
Posts: 5609
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by zullil » Sun May 05, 2019 10:56 am

todd wrote:
Sun May 05, 2019 4:12 am
I've done a lot of SF depth 50+ analyses from a variety of positions in the 6. h3 line over the last few years.

If your goal is simply to get SF (without contempt) to say 0.00 rather than having an easy position to play, then we can start from the following position (unless you think white should do something else)

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e5 7. Nde2 h5 8. g3 Be6 9. Be2 b5 10. O-O Nbd7
You meant 9. Bg2, with your line leading to

Karol Majewski
Posts: 94
Joined: Sat Jul 09, 2011 2:18 pm

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by Karol Majewski » Sun May 05, 2019 1:39 pm

Thx for your input, guys.
todd wrote:
Sun May 05, 2019 4:12 am

If your goal is simply to get SF (without contempt) to say 0.00 rather than having an easy position to play, then we can start from the following position (unless you think white should do something else)

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e5 7. Nde2 h5 8. g3 Be6 9. Be2 b5 10. O-O Nbd7
I remember 8...Be6 was played by Giri here. There are also games with 8...Nbd7 here. Maybe it was MVL? Anyway, I've looked at Be6 line and did not like the outcome. But now my Stockfish likes 8...Be7, which is what consistent with Louis' analysis:



2:51.16: +0.14/47 8...Be7 9.Bg2 b5 10.Nd5 Nxd5 11.Qxd5 Ra7 12.Be3 Rb7 13.O-O Nd7 14.Qd2 O-O 15.Rac1 Nf6 16.Rfd1 Be6 17.b3 Qb8 18.Nc3 Rc8 19.Nd5 Nxd5 20.exd5 Bf5 21.Qe2 Rc3 22.Bd2 Rxc2 23.Rxc2 Bxc2 24.Rc1 Bg6 25.Rc6 Ra7 26.h4 Ra8 27.Bf3 Qa7 28.Be3 Qd7 29.Bxh5 Qf5 30.Bg4 Qb1+ 31.Rc1 Qd3 32.Qxd3 Bxd3 33.Rc6 b4 34.h5

So, the main lines are:
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e5 7. Nde2 h5 8. g3 Be7
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e6 7. g4 h6 8. Bg2 Be7

I'll take a look a this 6...e6 7.g4 h6 stuff now...

jdart
Posts: 3816
Joined: Fri Mar 10, 2006 4:23 am
Location: http://www.arasanchess.org

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by jdart » Sun May 05, 2019 2:13 pm

7. Nb3 (after 6. .. e5) is also a serious line and not so easy to counter. Black's best is probably .. Be7 (Be6 is less good). Then there are two main lines:

8. g4 h6! 9. Be3 Be6!
and
8. Be3 Be6! 9. f4! Nc6! as in Stukopin-Dragun, Saint Louis 2018

Computer analysis can be helpful but it is also helpful to look at what the correspondence players do (I have Corr Database 2018 + recent ICCF games) and there is a Najdorf Powerbook available from Chessbase.

--Jon

User avatar
Eelco de Groot
Posts: 4148
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by Eelco de Groot » Mon May 06, 2019 1:49 am

Karol Majewski wrote:
Sun May 05, 2019 1:39 pm

So, the main lines are:
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e5 7. Nde2 h5 8. g3 Be7
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. h3 e6 7. g4 h6 8. Bg2 Be7

I'll take a look a this 6...e6 7.g4 h6 stuff now...
6...h6 (h6 and e6 can transpose as can other moves I think) seems a nice move to build a system repertoire around maybe (against humans or avoid computer bookmoves), because it seems to be not so much played. Just a casual observation. I had the same Stockfish Dev running as Louis I think, latest version 2019-05-04 Only because of default contempt 24 perhaps, evals came down to 0 rather quick. Don't think that means very much if SF does not really understand it, will just avoid blundering. Not saying it does not understand opening, I have no idea.


This was at the end, but not much changed for a long time evalwise (all zeroes or one miniscule small plus sometimes, or a minus even). Contempt 24, the default.



Engine: Stock Dev 20190504 (512 MB, one thread, my own TDM MingW compile)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

51 1522:32 0.00 1...e5 2.Nde2 b5 3.g4 b4 4.Nd5 Nxd5
5.Qxd5 Ra7 6.Be3 Be6 7.Qd2 Rd7
8.Qxb4 d5 9.Qa4 Be7 10.c3 Bg5
11.O-O-O Bxe3+ 12.fxe3 O-O 13.Bg2 dxe4
14.Qxe4 Qa5 (134.189.640.239) 1468

51 1522:32 0.00 1...e6 2.g4 Nfd7 3.h4 b5 4.a3 Be7
5.Be3 Nb6 6.g5 Bb7 7.Bd3 N8d7 8.f4 Qc7
9.Qe2 d5 10.h5 b4 11.axb4 Bxb4 12.e5 Nc4
13.Bxc4 dxc4 14.Rh3 Nb6 (134.189.640.239) 1468

51 1522:32 0.00 1...h6 2.g4 e6 3.Bg2 Be7 4.Qd3 Nc6
5.Nxc6 bxc6 6.e5 Nd5 7.exd6 Qxd6
8.O-O O-O 9.Na4 Qc7 10.c4 Nf4
11.Bxf4 Qxf4 12.Bxc6 Ra7 13.b3 Bd6
14.Qg3 Qxg3+ (134.189.640.239) 1468

51 1522:32 +0.09 1...b5 2.g4 Bb7 3.Bg2 Nc6 4.Nxc6 Bxc6
5.g5 Nd7 6.Be3 Rc8 7.Qd4 e5 8.Qd2 Nb6
9.b3 Be7 10.h4 O-O 11.O-O-O Qc7
12.Kb1 Bd7 13.Ne2 Rfd8 14.Bxb6 Qxb6 (134.189.640.239) 1468
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

jdart
Posts: 3816
Joined: Fri Mar 10, 2006 4:23 am
Location: http://www.arasanchess.org

Re: How to equalize in 6.h3 Sicilian Najdorf (Adams Attack)?

Post by jdart » Mon May 06, 2019 2:03 am

6. .. e6 7. g4 h6 has been tried by quite a few strong players and looks pretty reliable, although this is actually a transposition to the Keres Attack (B81), which Black should generally avoid. Usually White goes 8. Be3. One example of this line:


Post Reply