At best, white achieves a pawn-up situation after some exchanges, thus this can't possibly constitute the solution to the problem. However, if we can make the ..d5 for black impossible, we can still try to exploit the pin. Hence, Be6+! wins the exchange and the game.

The point is that white can *apparently* win the pinned rook on c4 -- just gang up on it with Rd4 or Rc1 until it falls. But black has a defence (after, say, 28.Rd4) in 28...d5 -- protecting the rook and also unpinning it.

But after a preliminary 28.Be6+ everything works, because ...d5 can be answered with check, eg 29.Rd4 d5 30.exd5+ Kf7 31.d6 Rxd6 32.Rxc4 Qd7 33.Rc7+ 1-0.

The right move for Black against this kind of rook penetration by White should be 34...Bf8, not Bh6. This way the black bishop isn't under attack after Rxh7, and it keeps an eye on the two important squares g7 and e7. After 35.Rxh7 Rd6 (or Qe5) the pair of rooks on the 7th may look scary, but White doesn't seem to have anything decisive.>

Ok lets take a look on how to counter your move,

35...Rd6
36. then the quiet move h3

What will be your move then? Qe5 or Rc6 or Re6 or any other else?

Rd4 was met with 28. ... d5 which lead to exchanges on c4 and a roughly even position. e5 was also met with 28. ... d5!, and after exf6, White's attack seemed to be stifled for the moment. I realized that if Black could play d5, he could greatly strengthen his position.

I kept that in mind when I analyzed Be6+. Initially, this was my least desirable candidate move because it gave up material for a seemingly meaningless check. Besides, the bishop prevented Rc8.

After 28. Be6+ Kxe6 (28. Be6+ Ke7?? 29. Bxc4 wins), it dawned on me that, with the Black's king on e6, d5 could not be played without incurring check (e.g. exd5+). It also meant that Black had no sufficient way of protecting his rook on c4. Suddenly, 28. Be6+ Kxe6 29. Rd4 (Rc1 transposes) d5 30. exd5+ lead to significant material loss for Black.

Okay, 28. Be6+ Kxe6 29. Rd4 (Rc1 transposes) Kd7 or Ke7 made sense. 29. ... Kd7 30. Rxc4 Qb7 31. Rfc1 d5 (nothing better?) 32. exd5 Qxd5 33. Qe2 was really bad for Black. 29. ... Ke7 30. Rxc4 Qb7 31. Rfc1 Rd7 32. Rc8 d5 33. Qa3 Kf7 looked winning for White. Apparently, Black thought so as well.

This game is a good example of flexible opening play on White's part. Nf3, c4 in response to c5 leaves all options open. It isn't until White's d4 in response to Nc6 that we have any idea where things are going. 4. ... g6 indicates a Sicilian Accelerated Fianchetto, though lesser played variations of the English are still possible.

By move 14, we are pretty much out of the opening. The position is even and traditional Sicilian style play commences. Although the material gains from their respective pawn forays are roughly equal, White succeeds in breaking up Black's king side yielding him a small, but tangible advantage. At this point, one other theme has emerged; White has firm control over the light squares.

With 23. ... exf6?, Black commits a small, but serious error. Perhaps he feared the rook pin and the fact that he would have had to play tenacious defense as a result of Bxf6, or maybe he wanted to play Bf8 protecting the isolated queen pawn. Regardless, the text move effectively shuts in Black's dark squared bishop since its difficult to imagine Black playing f5. 23. ... Bxf6 24. Qh6!? Rh8 25. Rd3 results in an awkward, difficult position for both players.

Beliavsky vs M Cebalo, 1998 (28.?)

White to play and win.

Material: Down a P with Bs of opposite color. The Black Kf7 has 4 legal moves. The White Qa2 pins Rc4 to Kf7. The White Rd3 and Rf1 both have semi-open files, Rf1 pinning Pf6 to Kf7. The White Bg4 can give check at e6. The White Kh1 is secure, but vulnerable to back-rank mates. The Black Qc7 must defend Rc4, so White should try to overburden Qc7.

Candidates (28.): Rd4, e5

<I can't see the reason for 29 ....Qc6! (with exclamation mark in your comment.Is it really a good move? Why not 29 .....Qc8 and if 30 d7 Qc7. 31 Bf3 Kf8>

I realize that this is old news, but my computer monitor died and I couldn't reply earlier.

I've looked through all the previous posts, including those of yours, and I don't much actual analysis of 28. e5 d5. It seems to me that after 29. Bf3 white has much the better game.

I won't go into deep variations either, since this puzzle is last week's, but it looks as if black must acquiesce to Bd5+ and a dangerous initiative for white at some point or else go in for lines that give him other tough problems.

Two blunders must be avoided. Obviously not 29 . . . fxe5 30. Bxd5 ++. Less obviously, not 29 . . . Qxe5 30. Rxd5 Rxd5 (virtually forced) 31. Qxc4 wins at least a rook. Other plausible 29th moves for black include Rc2, Rc1, and Kf8, as well as perhaps Rxb4. On . . . Rc2, 30. Bxd5+ followed by an appropriate queen move (a3 or xa5). On Rc1, white can either capture with check on d5 immediately or first play 30. Rxc1 Qxc1 31. Rd1. On 29 . . . Rb4 again 30. Bxd5+ and then either Qa3 or Qxa5; in any case the black rook seems misplaced on b4. On 29 . . . Kf8 simply 30. Qxa5 gives white a strong passed pawn and play also against the exposed black king. Example: 30 . . . Qxe5? 31. Qxc4 dxc4 31. Rxd8+ and white will double the rooks (if 31 . . . Kf7 the tempo check 32. Rd7+ will allow the doubling still without letting the black bishop into the game). In my opinion this is won for white.

The pin suggests 28.Rd4 but Black has 28... d5. Therefore, 28.Be6+, to capture on d5 with check, 28... Kxe6 (else drop the rook on c4) 29.Rd4:

A) 29... d5 30.exd5+ and 31.Rxc4 wins an exchange.

According to the computer, Black's actual, decisive mistake was 25...Rd8?, allowing 26. b4! Rxc4 27. Qa2 +- (+3.25 @ 32 ply, Stockfish 10).

Instead of 25...Rd8?, Black could have put up more resistance for practical drawing chances with 25...Bf8 26. Rh3 ± (+1.10 @ 33 ply, Stockfish 10).

Earlier, instead of 23...exf6? 24. Rd3 ± (+0.84 @ 35 ply, Stockfish 10), exposing a weak pawn on d6, Black could have kept the game near level with 23...Bxf6 24. Qh6 Rh8 ⩲ to = (+0.34 @ 33 ply, Stockfish 10).

After 28. ♗e6 ♔xe6 the defence d5 wouldn't work since e4xd5+ is a check!