MartinC wrote on 02/25/14 at 09:42:57:


I agree. As I  said, I wouldn't want to defend this in OTB play.

I checked my copy of Berg' GM Repertoire 15 yesterday and was kinda surprised that he doesn't even mention 12. Qg3. I guess this variation is dead since Roth - Kindermann (Vienna, 1996), but as far as I've seen it goes unmentioned. For anyone who hasn't seen it, here it is for pure pleasure:



edit: Ok, Berg has no need showing this move, because he recommends 11...Rf7 instead of 11...e5 (?! according to him), so he only proves his point by showing optimal play by white (in his view). Still talking about 11...e5 the game Roth - Kindermann is worth being posted here imo.

11...e5!? (Portisch 1958)

Very good story about this move given in Dvoretsky "Opening Preparation": Zlotnik recalled this idea 30 years later (!) to Dokhoian and his trainer Kishnev, as an alternative to 11..Rf7

12.Qh4 e4 13.gxf6 14.Qxf6 

A) 14...Qf8 15.Qxf8+ Kxf8 16.dxc5! dxf3 17.gxf3 (Kindermann/Dirr, and Watson). For me this variation is += or +/- for White: can lead to the positional pawn structure nightmare a3,c2,c3,f2,f3,h2 (6 "isolated" pawns) but otherwise White is slightly better.

B) 14...exf3 15.Rg1 Ng6 16.Qxd8+ Nxd8 17.gxf3 Kg7 18.dxc5 Ne6 19.Rd1 (19.ooo Kf7 20.c4)  Nxc5 20.Bxg6 hxg6 21.Rxd5 b6 22.Rdg5 
Not the only variation, but depending of position of Black pieces, some lines give little advantage

C) 14...cxd3 15.cxd3 Bf5!? (15..cxd4 16.Nxd4 Nxd4 17.Qxd4 Nc6 (17...Bf5 18.oo Nc6 19.Qe3 d4 transposes) 18.Qe3 d4 19.cxd4 Bf5 20.oo Qxd4 21.Rfe1 Rf8 (21....Qxd3 22.Qxd3 Bxd3 23.Rad1 Bf5 24.f3 += with the plan g4 and pawns advance on kingside wing; 21..Qxe3 22.Rxe3 Rd8 23.g4 Bg6 24.Rb1 +=) 22.Rab1 Qd7=



B)

Something interesting from reading Berg's volume 2 book on the Winaver, which contains a chunk on the 7 .. o-o Winaver. 8 Nf3 has never been meant to be critical, and still isn't. This has traditionally been because of a lot of extra sharp options black gets. What is interesting there is that this book is (I believe) the first reasonably modern text to examine a lot of blacks very sharp lines with a high powered computer/very strong GM etc.

Consqeuently perhaps notable how many of these ideas Berg doesn't really like - he doesn't really like 8 .. f5 9 ef Rxf6 10 Bg5 Nd7 or Qa5, and gives detailed reasons for the former.

The old Kindermann/Dirr idea of 8.. Qa5 ^ c4 seems to have survived intact. (If not credited to them Since played by other people etc it seems.). 

Black can also go 8.. Nbc6 of course, which is rather important due to the 8 Bd3 Nbc6 9 Nf3 move order. That then leads to the final alternative position which starts below.



Now here 11.. Rf7 is fine (and covered in some detail), but does seem like a tiny bit of a retreat back to 8.. f5 based ideas.

The 'fun' line starts with 11.. e5 when it seemingly comes down to evaluating this ending:



All known theory/from several games this. You can't analyse that out concretely and he doesn't try. He clearly doesn't like it terribly much for black though, while Watson did in PTF III. Any suggestions/ideas? Its a rather interesting position of course.

My (very humble) feelings, having played something like this three times (at a modest level), is that I'm not sure Berg is right to be worrying about blacks king - the pieces/black rook tend to get active enough that white has to take Q's off at some point - but I'm not hugely convinced about how attractive the resulting ending is for black.

It is relatively easy to see how whites kingside could end up somehow rolling over black, while its much less obvious how black can hope to win.
(The one win I did get was when white rather careleslly let me keep the d4 pawn alive. Otherwise a couple of games which held but with a bit of work involved.).

One thing I can remember not being sure about when analysing my win was whether black should actually even bother recapturing the pawn on d4. The white pawn on d4 really isn't doing anything terribly relevant to help white and it does make d5 a rather nice outpost for a piece/save a few tempi. Nothing that'd change the long term features of the position though.