Like <c4gambit> and <chesskidnate>, I went for 30...Rxe1 31. Bxf6 Rxf1+, when:

a) 32. Qxf1 d2 33. Qd1 (33. Bxc3 d1=Q is hopeless for White) Bxf6 is an easy win for Black because he simply defends the e2 pawn with his bishop and plays Rc5-c1.

First move 29...Rxe1 is easy to guess. After 30.Bxf6, I guessed 30...Rxf1+.

click for larger view

<Patriot> I prefer 30...♖xd4 to protect the Black d-pawn. If 30...♗xd4, then 31, ♕xd3. So, 30...♖xd4 is much better than 30...♗xd4

30...♖xd4, 31. ♖e8+ ♔g7, 32. gxf5 d2!, 33. ♕g3+ ♔h6

31. ♕xc3 ♖xf1+, 32. ♔g2 ♖f2+!! - see below

1) 33. ♔xf2 d1=♘+!, 34. ♔g3 ♘xc3, 35. ♗xc3 fxg4

2) 33. ♔h3 ♖f3+!, 34. ♕xf3 d1=♕

Part 1 of 2<<<>>>>

The players who produced today's puzzle are (near-) eponyms of two far-famed authors of centuries past: Anton Chekhov and E.T.A. Hoffmann. But, although the former is far better known, it seems that this time around, it is the latter who has won the laurels: Black's position is so strong, he can passively sacrifice his queen and then take his choice of winning methods.

Of course, no work is great without a great beginning, and this time it takes the form of

<29. ...Rxe1!<>>

This offers a queen, and if he is to have any chance at all, White will have to take it. To prove that, let's consider what happens if he doesn't.

<(a) 30. Rxe1?, Rxd4
31. Re8†, Kg7 >

There really isn't much to look at here. White will have to finagle to find so much as a pair of spite checks, and then reality falls upon him like Mr. Croup: all sharp-pointed bishops and skull-crushing advanced passed pawns and the agonies of dissolution.

<(b) 30. Qxd3?, Rxf1†> and either

<(b.1) 31. Kg2/Qxf1, Qxd4 > or

<(b.2) 31. Bg1, Rxd3 >

If anything, this is worse.

<(c) 30. Bxc3?, Qxc3 >

Again, White has nothing left to hope for.

This appears to exhaust the plausible "alternatives." Now we return to our main line.

<30. Bxf6, d2!<>>

Sharp to the point of insolence. Unless White finds a killing resource, Black will force in the pawn and simply smother him. I see three ideas worth examining.

<(1) 31. Bxc3?!, Rxf1†
32. Kg2...>

If 32. Qxf1, d1=Q; 33. Be1 (to avoid trading queens), fxg4 and White has no fight left in him.

<32. ...d1=Q >

But here too White has no threats. So much for this short line.

<(2) 31. Qh6...>

Now, this is more like it. If Black isn't careful, White can draw or even win.

Part 2 of 2<<<>>>>

<31. ...Rxf1†
32. Kg2, Bxf6>

This eliminates the mate threat.

<33. Qxf6, fxg4!>

With the rook on d5 now eying g5, White has one check on d8. Or he can get what he can.

<34. Qxf1, d1=Q >

And what is White to do? He has no checks, and if he avoids the exchange, Black begins checking; when that happens, White's king will soon emulate Charles I of England in 1649.

Now back to reality.

<(3) 31. Qxc3! ...<>>

Now the real fun begins. White is temporarily ahead on material, he controls the long diagonal leading into Black's kingside, and he threatens mate with 32. Qc8†. But of course one good resource deserves another.

<31. ...Rxf1†
32. Kg2, Rf2†!<>>

White has five choices, although some fail immediately.

<(3.1) 33. Kh1??, d1=Q†
34. Qe1, Qxe1#>

<(3.2) 33. Kg1?, d1=Q†
34. Kxf2, Rd2† >

White will soon have to part with the queen to avoid mate.

<(3.3) 33. Kxf2, d1=N† >

The point of 32. ...Rf2†: Now Black has an easily won ending after 34. Ke2 (not to the third rank because of rook forks), Nxc3; 35. Bxc3, fxg4, when the second player leads by the exchange and three pawns.

<(3.4) 33. Kh3, Rf3†!
34. Qxf3, d1=Q >

Thanks to the rook on d5, White's queen on f3 has no checks, and the first player is again confronted by a grim choice: Trade into a lost ending, or move the queen and let Black start his own checks.

This leaves one defense:

<(3.5) 33. Kg3, f4!<>>

What a quandary for White! We already know from line (3.3) what happens if 34. Kxf2, d1=N†, etc. We also know from variation (3.4) what happens on 34. Kh3, Rf3†!; 35. Qxf3, d1=Q, etc. This leaves one independent option, but we will soon see what it is worth.

<34. Kh4?, Rxh2†
35. Qh3, Rxh3 <>>

After 30. Bxf6

click for larger view

The game move (30...d2) is initially Fritzie's top pick. But after a minute or so on my medium-spec PC he replaces it with the move I preferred in human mode (30...Rxf1+). The point is that after 31. Qxf1 d2, black is going to lose most or all of his material trying to stop the d pawn from queening.

The game line (30...d2) also wins handily. There's not much between them (+5.7 for 30...Rxf1+ and +4.8 for 30...d2).

Back to the game line. After 31. Qh6

click for larger view

Either of 31...Bxf6 or 31...Rxf1+ wins for black. Fritzie is currently saying that both are mate in 15. Yeah, right.

After 32. Kg2

click for larger view

We have three winning moves: 32...Bxf6, 32...Rg1+ and 32...Rf2+. In each case, the idea is to either play Bxf6 or Rxg4 to neuter the white mating threat on g7.

After 33. Kf2

click for larger view

there are three and a half ways to win: 33...Bxf6, 33...Rxg4 (the game line) and 33...Rg2+ (followed by either d1=Q+ or Bxf6). There is also 33...Bxd4 which earns a +2.7. Winning, but not as forceful as the other lines.

Yes, that seems to work fine. After 33. Qxf6

click for larger view

White's last trick is the perpetual Qd8+ - Kg7 - Qg5+. Fritzie finds no fewer than 6 defences which bring home the point for black:

< it's definitely way too easy for a Sunday! >>

At this point, White's game is lost. The best he can hope for is B vs R+P endgame. 31. Qh6 Rxf1+ 32. Kg2 Rg1!+ or 31. Bxc3 Rxf1+ 32. Qxf1 (32. Kg2 d1=Q ) d1=Q 33. Qxd1 Rxd1+

White threatens 30.Bxf6, 30.Rxe8+ and possibly gxf5 to open the g-file.

The pawn on d3 suggests 29... Rxe1, to weaken White's first rank:

A) 30.Bxf6 Rxf1+

A.1) 31.Qxf1 d2

A.1.a) 32.Qc4 d1=Q+ 33.Kg2 Rd2+ with an extra rook and a winning attack.

A.1.b) 32.Bxc3 d1=Q 33.Qxd1 (33.Be1 Re5 - + [R+2P vs B]) 33... Rxd1+ 34.Kg2 fxg4 35.Bxa5 Ra1 36.Bb4 Rxa2+ - + [R+3P vs B].

A.2) 31.Kg2 Bxf6 32.Kxf1 d2 - +.

< Steve.Patzer: Does 31. ♕xc3 prolong the game? >

<31. ♕xc3 ♖xf1+, 32. ♔g2 ♖f2+!! - see below

1) 33. ♔xf2 d1=♘+!, 34. ♔g3 ♘xc3, 35. ♗xc3 fxg4

2) 33. ♔h3 ♖f3+!, 34. ♕xf3 d1=♕ >

Instead of 32...♖f2+, 32...♖g1+!! is much better - see below:

1) 33. ♔f2 d1=♘+!, 34. ♔xg1 ♘xc3, 35. ♗xc3 fxg4

2) 33. ♔h3 ♖c1! (preventing 34. ♕c8#), 34. ♕b3 (Δ ♕b8#) ♖d3+!!, 35. ♕xd3 ▢ fxg4+!, 36. ♔xg4 d1=♕+ (forcing the queen trade)

Sorry for my mistake, chess folks

34. Qb2 fxg4+!, 35. Kh4 Rc8, 36. Bh8 Kf8!, 37. Qg7+ Ke7, 38. Qxh6 d1=Q (White runs out of the queen checks)