The game began
1. e4 c5 2. d4 cxd4 3. Nf3 Nc6 4. c3 dxc3 5. Nxc3 g6 6. Bc4 d6 7. Qb3
Hitting f7, and encouraging ...e6, whereupon Black's Dragon formation "loses its shape", as Marc Esserman puts it in his book.
7...e6 8. O-O Nf6 9. Rd1 Qe7 $6 10. Bg5
10. Bf4 is even stronger, with the idea of encouraging Black to weaken d5 further with 10...e5 11. Bg5.
10... a6?
10... Bg7 was best, and then if 11. Nd5?! (11. Bf4) (11. Nb5) 11... exd5 12. exd5 O-O
13. dxc6 bxc6.
11. Nd5
11. Bd5! was even stronger, but sacrificing the knight is the second best move in the position.
11... exd5 12. exd5 Ne5
12... Bg7 13. dxc6 O-O is objectively best, but then White has attacking
chances and the better pawn structure in a position with level material, and
the black pawn on d6 will probably drop off.
13. Nxe5 Qxe5
13... dxe5 14. d6 gives White a winning attack.
14. Re1
14. f4! would have been even stronger, preventing Black's ...Ne4 idea, but a couple of lines have to be calculated accurately. The move played in the game is also winning for White, though.
14... Ne4
15. Rxe4! Qxe4 16. Qc3! Qe5
I saw 16... Bg7 17. Qxg7 Rf8 18. Bh6 Qe7 19. Qc3 renewing the threat of Re1.
16...f6 17. Re1 Qxe1+ 18. Qxe1+ Kf7 is Black's best chance, whereupon White has to
find 19.Bxf6! to get a winning position.
17. Re1 Bg7
17... Qxe1+ 18.Qxe1+ Kd7 gives White a choice of winning continuations. For example, 19. Bf6 Rg8 20. Qe3 b5 (20... Kc7 21. Qe8) 21. Qb6! forces mate.)
18. f4!
18. Rxe5+? Bxe5 followed by ...0-0 is probably alright for Black.
18...Qxe1+ 19. Qxe1+ Kf8 20. Qe7+ Kg8 21. Qe8+ Bf8
Unfortunately this is the point where I bottled it. 22.Bh6, which was my original intention, forces mate in four moves, but I somehow convinced myself that Black could wriggle out, and consequently played a "safer" move that allowed Black to wriggle out...
22. Be7?? Kg7! 23. Bxf8+?!
As is often the case after the psychological blow of missing a quick win, I rather went to
pieces. 23. Bh4 would still have left White with a strong, possibly winning,
attack.
23... Rxf8 24. Qe4?! Bf5 25. Qd4+ Kg8 26. Qb6 Rac8 27. Bb3?! Rfe8 and Black went on to win.
Unfortunately I have been forced to link to the game rather than embed a replayable game into the blog post, because at least at my end I seem to have triggered some sort of block that is stopping me from viewing replayable chess games in my own blog (but not other people's blogs).
http://view.chessbase.com/cbreader/2019/3/7/Game19546281.html
An agonising loss, but I do seem to have improved my attacking play a bit recently - a common past failing of mine has been to be too eager to "cash in" and missing the quiet moves (like 16.Qc3 and 18.f4 in the above game). I think had I reached the position at move 15 this time last year I would probably have been preoccupied with trying to regain the piece with f2-f3 and it is unlikely that I would have had the guts to play 15.Rxe4.
Having seen a number of AlphaZero-Stockfish matches recently, I think I may well have been inspired by AlphaZero's willingness to sacrifice pawns and even pieces and then quietly improve its position.
