Candidate Moves: ♖xe6, ♗xh7+, ♗h4, ♕e2

Seems that chessgames.com is back in schedule. After about 5 minutes of searching, I decided upon

13. ♖xe6 ♕xe6 14. ♗xh7+

forcing

14...♔h8

and after 15. ♕h5 I thought White had a clear win, but apparently Black has the saving move 15...♕h6 which stops the attack entirely.

So, I started off looking at 13.Bxh7 Kxh7. Now what? I now have 14.Qd3+, which takes me nowhere. I also considered moving the knight to clear the the way for Qh5+, but again, nowhere.

Fortunately, I didn't give up, nor did my faith in Bxh7 waiver (much). About the only other notable thing about the king on h7 is that it is in range of my knight, which isn't very useful at the moment because of the black queen, but can I deflect the queen in some useful way?

That's when 14.Rxe6! (finally) came to mind. It either deflects the queen allowing a royal knight fork, or it wins back a piece with 2 pawns to the good, since the rook itself is forking the Q+N.

pretty long continuation concidering its a medium puzzle sometimes you win in one move and sometimes more so I get confused how long I need to calculate in each puzzle....

Well, black has better ways to guard the knight than to play 14...e5 (which allows the white queen to hit d5).

Probably the best move for black at move 14 [diagram] is:

click for larger view

<13.?>

White:
a2, b2, c2, f2, g2, h2, Nc3, Nf3, Bd3, Bg3, Ra1, Re1, Qd1, Kg1 Black:
a6, b5, d5, e6, g7, h7, Nc6, Nd7, Bc8, Bc5, Ra8, Rf8, Qf6, Kg8 Material: =
Candidates: 13.Bxh7†, 13.Bh4, 13.Nxd5, 13.Rxe6, 13.Bxh7†

-ML-
No, I have not made a typo by posting Bxh7† twice in the candidate list: at first, it didn't lead to anything so I wanted to trap the Black Queen, which failed too. Then, Nxd5 comes to mind: does not work at all. Rxe6 comes close to the idea and gives the impression that that is not the most forcing line. Only after seeing the fork, I went back to:

<13.Bxh7†!> I will NOT vary with 13....Kh8. This loses a pawn for nothing, without a White attack as compensation. Analysing that is like replaying this game with a better move for Black. There are still two replies to look into though.

/A\
<13....Kf7> is objectively worse than Kxh7 as it does not capture the piece outright and the Queen has less moves. As:

<14.Rxe6!> is just as effective.

/AA\
<14....Qxe6 15.Ng5† Ke7 16.Nxe6 if Kxe6 17.Qxd5† >

/AB\
<14....Kxe6 15.Qxd5† Ke6 16.Re1† Kd8 (Nce5 17.Nxe5 Nxe5 18.Qxc5† ) 17.Bh4 > maybe not as clear as I state here, but I think Rxe6 is playable because of this: in the brackets variation the entire Black camp collapses.

/AC\
<14....Qd8> opposed to B's Qf5 which is a better square...

<15.Rxc6 > ...it's identical to B's main line in that White wins two pawns.

/B\
<13....Kxh7 14.Rxe6!> as stated before, this is a little less tricky AND less vicious as the King is not involved directly.

/BA\
<14....Qxe6 15.Ng5† Kg6 16.Nxe6 > without the Black King nearby this is just Q for R+B+2P in White's favor.

/BB\
<14....Qf5> better than d8, see A.

<15.Rxc6 > but the same outcome which is +2P.

Of course it is like replaying the game with a better move for black. Because solving the puzzle should not mean hoping for the opponent to make your inferior line look good by blundering. What sort of strategy is that? If you want to count on a blunder, then just move randomly, and when the opponent blunders, then you can feel justified.

It's pointless to consider the game line in this case, when all the success of that line for white is a gift from an inferior move by black.

Have you read my entire post? 13....Kh8 is my final variation (C) (actually, I see now that that HAD to be Kh8 but reads Kxh7), but I have not analysed it, as THAT would be like replaying the game. Without any clear win for White (not from what I've seen, at least) from that move on.

13. Rxe6 Qxe6
14. Bxh7+ Kxh7
15. Ng5+ - forking the queen and king.

White will have a queen and 2 pawns at the price of a bishop and rook. Also white will be able to trade the e6 knight (after the queen capture) for black's c5 bishop (because of the attack on the f8 rook). Next, it looks like the d5 pawn will fall. In any case, it is quite hopeless for black. (And who knows, perhaps even better opportunities pop up if this line gets played).

Of course, this is not all forced, so working backwards, lets say that black declines the white h7 bishop with:

14 ...Kh7 (the only square as f7 would also lead to a knight fork on the black king and queen). 15. Ng5

Now the black queen has the following moves: a) Qh6 b) Qf6 c) Qe7.

15...Qf6 or ...Qe7 allows 16. Qh5 with what looks like a very strong attack (e.g., the bishop can come out with a discovered check to g6).

15 ...Qh6 would seem to invite a move like 16. Qxd5. But then things start to get fuzzy for me. For instance 16 ...Nf6 and it is hard to see what white can do.

Going back even further, it doesn't look like black can decline the initial white rook sac (i.e., 13 Rxe6). Moving the queen away allows white to capture the c6 knight.

So I am betting that 13. Rxe6 is the correct initial move. Perhaps with 15 ...Qh6, white should play 16. Bf4 (instead of Qxd5). This contains the threat of playing Nf7+ with a discovered attack on the queen. Moving the black queen over to f6 again invites the white queen to move to h5. And moving the queen down to h4 looks uncomfortable.