30. ♗xg6 fxg6 31. ♕xg6 ♗g7 32. ♗e5 ♖h7 33. ♕xd6 ♕xd6 34 ♗xd6 ♗xc3 and the game is =

30. Bxg6 screams for attention. The g5 pawn is only defended once and is twice attacked. Strip away the f and g pawns and the black king is open to attack.

But like <lost in space> I wanted to cash in straight away with 31. Qxg6 Bg7 32. Be5 Rh7. That gives white an attacking position with moves like Be5, Re3, Rd3, Bxd6, bg5, Qe6+ to think about. We have to guard against Qh4, but other than that we are free to attack. I would rather be white in the resulting position, but there is nothing conclusive.

But I didn't spot the sublime 31. Re6. I suppose the lesson here is that a mating combination can include a threatening move - instead of the all-too obvious captures and checks.

Well, let's start with the obvious:

30 Bxg6

The reason that I say this is obvious is that black can't simply recapture with 30...fxg6 due to 31 Qxg6+ Bg7 32 Qxd6 and white has recovered his piece and is now up a pawn.

Since black must have seen this line when making his last move, then black must have something in mind to deal with this threat. I think black would try

30...Rh4

Now a normal discovered check is met with 31...Rxg4, so I expect white to try

31 Bxf7+

This is a double check, so the black king must move

31...Kxf7 32 Qe6+ Kg7 33 Be5+ Kh7 34 Bf6

Forking the rook and queen.

Wow, this line is so bad for black that I'm beginning to doubt the 30...Rh4 move.

Well, I'm sure that there's a LOT more here that I'm not seeing, but it's creeping up on midnight and I'm tapped out mentally.

Forking the rook and queen.>

Ahem, 34 ... Rh6 draws:

click for larger view

However, 30 Bxg6 Rh4 31 Bxf7+ Kxf7 32 Qe6+ Kg7 33 Rd3! and it's curtains for the black:

click for larger view

1. (+4.23) 37. f4 Rxe6 38. Rxe6 Re8 39. f5 Qa7 40. Kg2 Qf7...

2. (+2.58) 37. Rxg6 Qxg4 38. Re7 Kh8 39. Rgxg7 Re8 40. Rh7 Kg8 ...

I saw that, too and thought 31. Bxf7+ Kxf7 32. Qe6+ Kg7 33. Be5+ is more than enough.>

I wanted to believe that 33. Be5+ was enough, too, but I simply couldn't see it.

Forced mate. Something like:

Fritz 9 agrees with 30. Bxg6 as the first move, but wants to defend with 30. ... Bg7 which it evaluates an advantage to white of 0.88. Nice to have, but not a game-winner.

On 30. ... fxg6, Fritz totally ignores the "obvious" 31. Qxg6 and goes straight for 31. Re6. The evaluation rises to 1.78 with lines such as 31. ... Bg7 32. Rxd6 Qh4 33. Qxh4 Rxh4 34. g3.

I am afraid that the tempting 31. Qxg6 gets an almost instant evaluation of =(0.00). Sitting on infinite analysis for a few minutes does nothing to shift the evaluation, although the fan in my computer does make a wonderfully loud whirring noise as the processor has to work hard for its living.

Dzechiel's line seems to fizzle out. Fritz gives 31. ... Bg7 32. Qxd6 Qxd6 33. Bxd6 Bxc3 with an evaluation of =(-0.25).

My line with Be5 fares no better, with Fritz refusing to budge from =(0.00). Black defends with Rh7 and at some point white has to do something about the fact that he is a piece down. Exchanging pieces gets to an endgame where white has passed f and g pawns and black has a passed d pawn.

So I guess that 30. Bxg6 is only strong if you follow up with Re6. Otherwise, it leads to an equal game. Mind you, as white starts the puzzle position a pawn down, you could decide that an equal game is worth having. Even with Re6, the advantage is not overwhelming.

....f7x[B]g6? 31. Qxg6+ Bg7 (forced) 32.Be5 Rh7 33. Re6 Ne8 and its gone off my radar - all too fuzzy.

However Judit, the giant killer, might have chosen the more strategic 30. Be5 Rh4 31.Qg3 since neither Ne4 nor Nf5 are playable so if Ne8 perhaps now 32.Bxg6 works.

Material: B for N+P. The most striking feature of the position is how the White Rd1 and Re1 dominate the center. The Rd1 is on the same file as 3 Black units: Pd5, Nd6, and Qd8. Because Bf4 attacks Nd6, removal of Pd5 would burden both Qd8 and Bf8 with protecting the Black Nd6. Also notable is the White B-pair. In concert with the White Qg4, Bc2 attacks Pg6, an obvious sacrificial target, because only Pf7 protects it. Moreover, the Black K-side has dark-square weaknesses, which Bf4 can exploit at e5. Finally, the White Qg4 pins Pg6 to Kg8. Overall, the harmonious White development is well worth a P and easily able to support the obvious sacrificial strike with 30.Bxg6. After removal of Pf7, the square e6 will be useful, particularly to Re1 and Qg4.

It seems this week's theme is "obvious first moves".

Step up, step up! Everyone's a winner!

Candidates (30.): Bxg6

30.Bxg6

Because the alternative is a losing position (inferior development and an open K facing a B pair), Black must accept the sacrifice of Bg6:

30…fxg6 31.Qxg6+ Bg7

Now, the next move is far from obvious. The obvious candidate 32.Bxd6 Rh6 puts White in dire straits; and neither 32.Qxd6 Qxd6 33.Bxd6 nor 32.Qe6+ Nf7 do the initial position justice.

As <dzechiel> is fond of saying: Rewind!

The obvious capture 31.Qxg6+ is an error. Instead, let's delay gratification and use e6 to bring more pieces to the front.

31.Re6 (threatening 32.Rxg6+ or 32.Qxg6+, then 33.Rxd6)

Black cannot defend Pg6:

31…Rh6 32.Bxh6 Bxh6 [else, drop an exchange]

33.Qxg6+ Bg7 [else, down a P] 34.Rxd6

Black cannot "counterattack" (31…Qh5 32.Rxg6+ and the end is near), so the only hope is to move Nd6. Of the 2 feasible moves,

32…Nf7 33.Rxg6+ Kh7 [Bg7 34.Rxg7+] 34.Qh5+

quickly surrenders the sacrificed material with interest. Thus,

32...Ne8 33.Rxg6+ Bg7

[33…K moves lead to mate]

[33…Ng7 34.Qe6+ Kh7 35.Rh6#]

34.Qe6+ (threatening a R lift with 35.Rd3)

34…Kf8 35.Rxg7 Kxg7 [Nxg7 36.Bd6+] [else, mate]

36.Rd3 (threatening 37.Rg3+)

36…Nf6 [Qf6 37.Be5] 37.Rg3+ Kf8 38.Bd6+

and White must surrender Qd8.