I sometimes see 3...h6 when I play the 3.Nd2 Tarrasch.  It makes no sense.  If black thinks that a move like 3...h6 makes sense, why not play it immediately in response to 1.e4?  All black is doing is trying to take white out of his French book.

In fact, I already played 4.a3 last year against Uberdeker. Its mockery value should not be underestimated.
However, I can't say if the Kosteniuk game followed our path because I don't remember Uberdeker's treatment (it was blitz).

I don't remember which game, but 3...h6 occurred in the recent (2006) Women's FIDE KO Championships and White responded with a very early a3.  I'll try to find the game.  However, I agree with others that this line is probably not in any way theoretically critical.  White won a fairly unremarkable game. 


If you're interested, I'll try to dig it up.  But as a French player, I'd hate to be pushing h6 without a specific reason.  I don't see a good reason for such a time waster on move 3 unless 3...Nf6 4.Bg5 scares Black so much.  Then maybe Black should play 1...c5 instead.

Hello Uberdeker,
The main idea for White to reply a Winawer strategy (Bb4 move) is to use Steinitz's idea of an exchange exd5.
It was known as the "refutation" of the Winawer until Nimzowitsch improves Black  strategy around 1920.
I will use precisely the classical move order to reach the position:
1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.exd5 (Steinitz) exd5 5.Nf3 h6 (Paulsen-Lichtenstein,Berlin 1864 ; nowadays Ne7/Nc6/Bg4/Nf6 are the moves ... I use only this game as an example of strange move for Black) 6.Bd3 Nf6 (6..Bg4 7.h3) and here is the position reached after 3..h6 4.Nf3 Nf6 5.Bd3 Bb4 6.exd5! (6.e5 is an interesting line leading to a gambit and need some deep analysis 6...Ne4 7.oo Nxc3 8.bxc3 Bxc3 9.Rb1 Bruzon-Arencibia,La Havanne 2004, Horvath-Borgio,Zemplinska Sirava 2004) exd5 7.oo oo 8.Ne2! (for me it's "simple chess" and I don't really trust White's Ne5/h3 as in Hautot-Legky,France N1 2005) Steinitz-Pitschel,Vienne 1873 ; Blackburne-Wood,Gread Brtiain simul 1871

About 4..Nc6!?: it's Eingorn's improvment over 4..Nf6 played earlier by Bricard. 4...Nf6. If White plays 5.e5 then it's variation given by Janev-Bricard,Toulon 1999: 5...Ne4 6.Nxe4 (6.Bd3 Bb4 is the gambit given before) dxe4 and now Black has a playable game because White must loose tempi to withdraw his knight on f3.

If White plays 5.Bd3 then it's too late for Black to play Nc6, simply because of 6.e5 attacking the knight on f6. Other moves: 5..dxe4?! (give up the center) Nxe4 ; 5...c5 6.exd5 exd5 7.oo  Stavast-Van Rin,Groningue 2003 ; 5...Be7 6.oo ; 5..Bb4 Jovanovic-Kovacevic,Zadar 2003 ; Bruzon-Arencibia,La Havanne 2004 6.exd5!

I add that exd5 is maybe the right answer in the line: 4.Bd3 Nf6 (instead of Eingorn Nc6) 5.exd5! exd5 6.Nge2 Almira Schripenko-Bricard,Montpellier 1988 (Wonderer on chesspublishing forum) or 6.Nf3

Dear Dom,

I don't understand your suggestion of 4. Ktf3 Ktc6. The whole idea of 4. Bd3 Ktc6 is to hit the unprotected -d4 pawn, coaxing White into 5. Ktf3, then chasing the Bd3 with 5. ...Ktb4 so that after 6. Be2 he can continue 6. ...Ktf6 ; 7. e5 Kte4
Now with the Kt allready on -f3, surely the logical course is 4. ...Ktf6. Now you have suggested 5. Bd3 c5 ; 6. 0-0, which looks reasonable enough for Black, but what of 5. ...Bb4  ?

                                                                                        Regards,
                                                                                             Hubert

Hello Mnb&Uberdeker,

Yes, maybe Bricard's variation is more appropriate than Eingorn's variation, but naming variations is difficult ("the first player playing the move ? the first strong player playing the move ? the first player building an opening system with it ?" and so on).

I add that I had the luck to listen Bricard in a training lesson in Toulouse few years ago, about the variation. He explained the ideas and one main idea is "h6 is a waiting move". Eingorn used Bricard's idea and developped later a system with in, and specially with the Nc6 move. The idea is, if White plays an "universal system", then White will not choose immediatly between closing or opening the center (e5 or exd5), then Nf3-Bd3 are the moves. But after Bd3 then Black has Nc6 hitting the bishop on d3, and if White wants to keep his very-good bishop, then one tempo must be lost.

Maybe I repeat the analyse in NiC, I don't know (because I have not read the paper about the variation).

Now, about the sub-lines...

A) Maybe it's a matter of taste, but c6 is another pawn move and White after the exchange exd5-exd5, two tempi more for the development and I clearly prefers White in such position. The idea of Bb4 is to win the good square "e4" and to develop quickly (the same idea exists in the Wiawer-exchange without h3).

B) The two-knights's variation is one sub-variation of the French where White delays the d4 pawn move. It's useful to know it for White players playing Nf3 or wanting to avoid some variations (White can choose later between the exchange or a closed game, and avoid the Winawer);
1.e4 e6 2.Nf3 d5 3.Nc3 two-knights variation (Taulbut) Nf6 4.e5 Nfd7 5.d4 c5 6.dxc5 Nc6 7.Bf4 Bxc5 8.Bd3 f6! (8...h6?! an error of a world champion, since now Black is weak on light squares and cannot play f6 easily. Gufeld-Spassky-Leningrad 1960)

C) Versus Kf6, White plays the universal system Bd3...but I admit Black can play such move. Without deep analysis, I will play with White: 5.Bd3 c5 (5...Nc6 6.e5 Nfd7 7.oo and now the idea Nb4 doesn't work 7..Nb4 8.Be2 c5 9.Nb5! a6 10.Nd6+ Bxd6 11.exd6 += ; 5...dxe4 is a kind of Rubinstein with the useless(?) move h6) 6.oo.

If 4...Nc6!? 5.e5 then Black has no immediate counterplay but an easy and quick development for his pieces: Bb4-Nge7-oo. The attack against the space advantage is done later with f6 or Ng6-Nce7

E)  After 6.Qxg7 Rg8 (6...Rh7 7.Qg3 dxe4 8.Bf4) 7.Qxh6 Nxe4 8.Ne2 and I don't see Black counterplay in the center, and White now has a dangerous h pawn in the future.

F) In some lines in the Rubinstein, Black can exchange more pieces because White has played Bg5. Here, the pawn move a3 has been played, and White can keep his space advantage or control: 4..dxe4 5.Nxe4 Nf6 6.Bd3 Nxe4 7.Bxe4 Nd7 8.Nf3 Nf6 9.Bd3 c5 10.Be3 or (Fort Knox's variation) 6...Bd7  7.Nf3  Bc6 8.Qe2 Bxe4 9.Bxe4 Qxe4 10.c6 oo

A : The Q Bishop has much more scope than in many other variations of the French. And here it is unopposed.
I suppose 10. Qe2+ should be met by 10. ...Be6 and 11. Qe5 c6

Dear Dom,

Thanks for your help. Your variations are interesting. However there are a few points which I haven't quite understood or on which I disagree :

line A) 5. Bf4 Bb4 looks a bit illogical, since the idea of 4. ed in the Winaver is precisely to demonstrate that the Bishop is misplaced on -b4 (although Black should manage to equalise without too much difficulty there). If 6. Ktb5 needs to be prevented, 5. ...c6 is probably better.

line B) What's the Two Knights' variation?

line C) 4. ...Ktf6 is the consistent move, taking advantage of White's Ktf3. 4. ...Ktc6 is met simply by 5. e5 and Black has no counterplay

line E) Again ...Ktc6 achieves little and you end up playing a Rubinstein where Be3 is probably worth more than ...h6. You imply that 4. ...Bb4 is met strongly by 5. Qg4, but 5. ...Ktf6 ;
6. Qxg7 Rg8 ; 7. Qxh6 Ktxe4 looks good for Black to me. What was your idea?

line F) Here however, the Rubinstein option 4. ...de is more attractive, since ...h6 is useful in preventing a  future Bg5 or Ktg5 and a3 serves no real purpose.

I don't think Grigoriants-Cosma is so thankless for Black. That bishop pair may come in handy at some point.

                                                                     Regards,
                                                                           UD

P.S. You're right MNb, 3. ...h6 was first played by French IM Bricard, but Eingorn's treatment is generally more dynamic and he's had success with it rather a high level.