What say you after 26. Nxd5 cxd5 27. Rc1 Re4 28. Rc7 Qb5 29. h4 Rae8 30. fxg6 fxg6 31. h5 R8e7? The engines seem to like it.

<26.Nxd5 cxd5 27.Raf1 Re4 28.h3!? Qb5 29.Kh2!? Be7 30.fxg6 fxg6 31.Qh6 Rh4 32.Qc1 Re4 33.Qc7 Re8

At the moment I can't decide between R1f2, b3 or Qa7/Qb7 here (maybe we should play all 3!). The idea of this line is simple - improve our position slowly as Black can't do very much. e.g. 34.R1f2 Re6 35.b3 Re1 36.Qa7.>

<whiteshark> says:
Last thought before bedtime:
What makes the difference of Qd3 in your line <26...cxd5 27. Qd3 Re4 28. Raf1> vs our mainline 27. Raf1 ? hmm

Maybe it's that black Re4 is not necessary, e.g.

26...cxd5 27. Qd3 Qe7 28. Kh1 Bh6 29. Raf1 Qe2 30. Qb3 Qc4

However, I think with the following White plan, White's assets can be realised fully - i.e. the 2-1 pawn majority on the queenside, and the bishop on b6:

Nxd5 cxd5
Qd3! Re4

(Q-d3 has the dual purpose of not only pressurising a6, but also supporting b4-b5)

Raf1 (to help consolidate the White king)
Bg7

h3 (further consolidation of the White king position) Qb7

Kh2 (consolidation further)

Rae8

and now and only now b4:

<kingscrusher> says: Whiteshark, the idea of Qd3 immediately is very simple - stop the blockade of the b-pawn by stamping out the possibility of black playing Qb5. The 2-1 pawn majority needs to be realised.

<thorsson> says:

2. What's the difference between 27.Qd3 Re4 28.Raf1 and 27.Raf1 Re4 28.Qd3? Well in the first Black doesn't have to play Re4, and in the second Black might play 27...Qb5. Let's take the latter first. 28.fxg6 fxg6 29.Qf4 looks good at first sight - does Black have anything better than retreating the Queen back whence she came? OK what about 27.Qd3? 27...Qe7 28.Raf1 Bh6 looks better than playing Re4.

I therefore suggest that we keep Qd3 in mind for next move.

<weisyschwarz>: The Qd3 idea looks interesting, and deserving of a forum. Playing both sides of the board is what grandmasters do. If we can incorporate elements from both Thorrson's and Kingscrusher's lines, it may go well. I like a possible Qb3 after fxg6 and b4.

<safar> says: After 26...cxd5 27. Qd3! if the fR moves to c8, then white plays aRf1 to put pressure on f7. Black's aR is tied down to defend the a pawn. If Black does not play the rook to c8 then White plays aRc1, followed by fRc3 and then Rc7. There is also the option of Qb3 to put pressure on the pawn on d6. I think that White is better here.

<izimbra> says:
Yes, but just to be clear I was discussing the motivation for getting in a Qd3 prior to blacks Qb5 - say after Raf1 R34 (there are factors against it, but this is the positive point). And I was saying that it makes the Qb5 followup for black, but black can still position the queen on the decent square c8, and go from there to c4. So this positive aspect of Qd3 may not be as big as it first appears.

<themadhair> says:
What say you after 26. Nxd5 cxd5 27. Rc1 Re4 28. Rc7 Qb5 29. h4 Rae8 30. fxg6 fxg6 31. h5 R8e7? The engines seem to like it.

<Willem Wallekers> says: The next decision point is after 26 ... cxd5. 27. fxg6 and 27. Rc1 are serious candidates.

<thorsson> replies to the above:

<domdaniel> says: I took a general look earlier at the idea of Rc1 and switching to the c-file (I'd suggested it in a forum and thought I should check it out), but I couldn't make it work. Without going into too many concrete lines, these were some of the problems: (1) Black plays ...Qb5 (threat ...Re2) and ...Rac8. White rook penetration via c7 is preventable.

(2) After ...Rac8, we can either exchange one pair of rooks, or scurry back to the f-file with loss of tempo. While several people have argued that exchanges favour us, the exchange of one pair of rooks is often the exception. One reason is that we're currently closer to doubling rooks on the f-file than black is to any equivalent maneuver. The 'rook pair' is one of our temporary advantages, not to be thrown away.

(3) In lines where we play fxg6 - which seems almost inevitable - it becomes clear that our rooks belong on the f-file. Again, getting there from c1 loses tempo.

(4) There might be a possible finesse - 'sacrificing' tempo to induce Black to do the same - in a line like 26.Nxd5 cxd5 27.Rc1 Rac8 28.Rcf1 Qb5 29.fxg6 fxg6 30.Qf4 -- although here I think we've sacrificed more with the two-step journey to f1, Black's rook is well placed on c8, and his a6 pawn is not weak. We have kingside threats, but nothing we can't get via 27.Raf1. The c-file feint arguably distracts black from the ...Re4 idea, but he gets plenty of compensation.

(5) Avoiding fxg6 and trying to operate entirely in the c-file, eg 26.Nxd5 cxd5 27.Rc1 Rac8 28.R3c3? simply gives Black time to play ...Qxf5 after chopping rooks. Finally, once rooks are opposed on the c-file we have to act - either exchanging them or moving the Rc1 away. I just tried this line on Fritz: it opted for Rc1-f1 rather than Rxc8.

The idea of 27.Rac1 will probably turn up a few times over the next couple of days. But I really think it's a non-runner, despite the optical attraction of the c-file.

(0.00) 28.Raf1 Qb7 29.Rh3 Bg5 30.Rg3 Qe7 31.Kh1 Bh6
(0.00) 28.h3 Qb7 29.Raf1 Rac8 30.Kh2 Re4 31.fxg6 fxg6

<domdaniel> says: 27.Rc1 Rac8 28.fxg6 Rxc1+ 29.Qxc1 fxg6 30.Qf1 - which is pretty promising for White, I think, although the engines don't rate it too highly. But it's better than the other Rc1 lines I had looked at. I suspect the low engine eval is due to a misreading of the bishop ending if all the heavy pieces come off, eg in a line like 30.Qf1 Qb7 31.Rc3 Re7 32.Qc1 followed by Rc7 and mass exchanges.

30.Qf1 puts black on the defensive immediately re the a6-pawn, and threatens Rf3-c3-c7 with mating threats. It also makes b4 easier, and black can't play ...Qb5 with our queen on f1. So it answers the brief of working on both sides of the board.

But black can deviate in a couple of ways. Mainly, he doesn't have to play 28...Rxc1+. After 28...fxg6, we're back in a line like the ones I considered earlier, where playing 29.Rxc8 achieves little and moving the c1-rook again wastes tempo.

<thorsson> says: After 27.Rc1 Rac8 28.fxg6 Rxc1+ 29.Qxc1 fxg6 let's just arrange a swap of the remaining Rooks and see what happens. 30.Re3 Rxe3 31.Qxe3 and Black has the tactical shot Qf5, so let's sort that out first, 30.h3 Bg7 31.Re3 Rxe3 32.Qxe3 Bf6 33.b4 Kg7 (not Black's most dynamic defense, but there's a point to be made) 34.Qe2 Qc8 35.b5 axb5 36.Qxb5 (could we ask for more?) 36...Qc1+ 37.Qf1 Qe3+ 38.Qf2 Qc1+ 39.Kh2 Bg5.

Not exhaustive by any means, but I think it ought to be clear that the Black Queen can hold the position on our own if we just go for a Rook exchange.

<whiteshark> says: 27.Rc1 Re4 28. Rc7 Qe8 29. fxg6 fxg6 30. h3 Re1+ 31. Rf1 Rxf1+ 32. Kxf1 Qe4 33. Qe2 Qf5+

sub-lines:
- 31. Kf2 Re4 32. Rcf7 Be7! 33. Qh6 Qxf7 34. Rxf7 Kxf7 35. Qxh7+ Kf6 (unclear) - 31. Kf2 Re4 32. Rff7 Be7! 33. Rf3 Bf8

<Artar1> says: have taken a look at 27.Rc1, and while it's not a bad move, it does not blow Black off the board:

27.Rc1 Qe7 28.h3 Rac8 29.Kh2 Qe4 30.Rff1 Bg7 31.Rce1 Qh4 32.Rxe8+ Rxe8 33.fxg6 fxg6 34.Qf2 Qe7 35.b4 <[35.Qf3 Qe6 36.Qd3 Qc8 37.Qd2 Qc4 38.Rc1 Qb5 39.Qg5 Kh8 40.Rc7 Qf1 ; 35.Qc2 Qd7 36.Qd3 Qc8 37.Qd2 Qc4 38.Rc1 Qb5 39.Qg5 Kh8 40.Rc7 Qf1 ]> 35...Qe6 36.Qg3 Qe2 37.Qf3 Qe6 38.Qd3 Ra8 39.Qb3

Each move was done to 18-ply using Fritz 8 on a 2 GHz machine over 6 hours. (Not exactly blitz chess.)