Toga declares the game variation best play (value 3.03 at 10 ply), and the alternative 27...Rg7 permitting the skewer of Rc8 as grossly inferior (value 12.25 at 10 ply). (I did not see any need to let the computer run longer than 10 ply.)

The plan behind the combination is not overly complicated. If White can divert the Bishop from the defense of b3 and cover the c2 escape square, then he can play Rb4 trapping the Queen. However, like many good plans, the difficulty in making it work is in putting together and executing the details -- or, in this case, finding the follow-up moves to make the combination work.

The first objective of Speelman's plan involves diverting the Bishop, which the Exchange sacrifice 24. Rxd7!! accomplishes after 24...Rxd7 25. exf6 Bxf6 26. Bxf6 gxf6. The second and more difficult goal, involves creating a threat against the King so White can gain time to cover the c2 escape square and then play the winning Rb4. Speelman accomplishes this with the clever follow-up move 27. Rh4! (not 27. Rg4+? Rg2 28. Rh4 =), which forces 27...Rf8 and allows 28. Qh8! Kf2 29. Qh7+ Ke8 30. Qg6+ Kd8 31. Rb4! .

Now, the objectives are achieved and the plan is complete, as the Black Queen is trapped and White wins!

P.S.: <johnlspuge> Look's like we came up with the same idea in our almost simultaneous posts. Thanks, as always, for your concise summary.

I read your deleted preliminary version, so I am not going to take complete credit for my prescience. To reciprocate, thanks for your complete synthesis. I often use your analyses offline.

I am not sure where <Once> is, but in his absence, I will steal a general observation originally his and express it less eloquently (no doubt) than he would.

What he didn't mention was that he was looking at a tic-tac-toe board with X in all 4 corners, O in every other square with a blank middle square.

Both 27 Rg4+ and 27 Qg6+ are stopped by 27… Rg7.

and white is a full pawn up :)

Candidates: 24. exf6, 24. Bxf6

The usual Mom-and-Pop Toga II 1.3.1 evaluation gives

[ply 15/57 time 05:18 value +0.36]

24.<Bh6> Nxe5 25.Rxd8+ Rxd8 26.Rxe5 fxe5 27.Qg6 Bf8 28.Ne4 Kh8 29.Ng5 Qd3 30.Nf7+ Kg8 31.Qxe6 Qb1+ 32.Kh2 Ra8 33.Be3 Kh7 34.Ng5+ Kh8 35.Qxe5 Qc2 36.Ne6 Kg8

Your move pulls but does not rock.

My chessforum has detailed instructions on downloading and using Toga, which is <free>. As an oracle, Toga is a useful tool for answering chess questions like yours, because they come up everywhere, not just on chessgames.com.

As an aside, you might improve faster if you focus on why a GM chose the game line, rather than on why the game line did not follow your move.

Keeping a very low profile because I totally and utterly failed to spot the solution today!

Had I spotted it, this is the sort of logic that might have led me there. I want to play Rb4, but the bishop on e7 stops me. To get rid of the bishop on e7, I need to play ef. But I can't do that straight away because the d7 knight covers f6. So first I have to eliminate the d7 knight.

And then to make sure that Rb4 truly traps the queen, I need to get my queen onto the long diagonal to cover c2.

I opened the CG website last night, after being mentally exhausted bysolvinga "hard" level sudoku. Thought Rh4 was winning (well, I had many lines in mind, but they all dealt with Kingside attack and had nothing to do with trapping the Queen. That's how the idea of Bh6 came to me.

I saw that the first three moves looked promising, but terminated my analysis there, not seeing anything in Qg6+ Rg7. I would never have imagined that an attack on the king culminating with Qg6+ would leave the queen stranded on the other end of the board!