After achieving my first IM norm at Queenstown in January last year, I came close to achieving my second through 2012 in a couple of tournaments, but couldn't quite manage it. I had been playing well in this Australian Open, and I could achieve the required rating performance with a win or draw as White against Andrew in Round 9.
Ikeda (2345) - Brown (2261), Australian Open (9.4), 11.01.13
1. e4 c5 2. Nf3 Nc6
Preparing for the game on the previous night, I was looking at a line in the Open Sicilian against the Accelerated Dragon which goes into an equal R+B+5P endgame. It seemed perfectly safe for a draw (and White scores only wins and draws in the database with it) which would guarantee the norm, but discussing my prep with Irene Sukandar (also playing at the tournament), she suggested an entirely different line. I felt good playing this tournament, and a part of me wanted to play for a win. What the 'correct decision' might be is difficult to say in a situation like this, but I decided on the latter variation this time.
3. Bb5 g6 4. O-O Bg7 5. c3 Nf6 6. d4
A line involving a pawn sacrifice, which Andrew had not seen before - definitely not a pleasant line to work things out against at the board.
...Nxe4 (most common and also critical is 6...cxd4) 7. d5 Nd6 (the other options are 7...Na5 and 7...Nb8, where White also retains the initiative) 8. Bd3 Ne5 (8...Na5 or 8...c4 are also playable) 9. Nxe5 Bxe5 10. Re1 Bg7 (10...Bf6 11.Bh6 and White has lasting pressure) 11. Bf4 (more promising may be 11.Qe2, as Vachier-Lagrave played in 2009, or 11.Bg5)
...Qb6?! (11...O-O was best, not fearing 12.Bxd6 exd6 or 12.Qe2 because of 12...e5! 13.dxe6 [13.Bxe5 Re8] ...dxe6) 12. a4?! (12.Nd2! developing would give White a sizeable advantage - 12.Qd2 or 12.Na3 were also better than the text) ...O-O (the immediate 12...c4 was also playable - the problem with 12.a4 is that attacking the queen with a5 doesn't achieve much, as she can still capture on b2 with tempo) 13. Rxe7 c4 (I think we both thought this better than the immediate 13...Qxb2 because of 14.Bxd6 Qxa1 15.Qb3 threatening Re7-e2-a2, but Black can escape with 15...a6! 16.Re2 c4! 17.Bxc4 b5! 18.axb5 axb5 19.Bxb5 Bxc3)
14. Bf1 Qxb2 15. Bxd6 Qxa1 16. Bxc4 b5
In our earlier calculations 16...Bxc3? 17.Nxc3 Qxc3 seemed weak because of a move like 18. Re4 and Black's position is a sorry sight (in fact it is close to equal after 18...b5! 19.axb5 Bb7), but we found later that there is an elegant win with 18.Re3! Qxc4 19.Bxf8 Kxf8 20.Qa1!, which deserves a separate diagram below:
|Variation: 16...Bxc3 17.Nxc3 Qxc3 18.Re3 Qxc4 19.Bxf8 Kxf8 20.Qa1 1-0|
Not liking the prospects of my c4-bishop, I played 20. d6?! but a more restrained 20.Qd3 or 20.g3 would have been stronger. ...Rb8? missing the chance to equalise with 20...Qb6 21.Qb3 Kg7 22.Bxf7, when the computer suggests the inhuman 22...Qc5!? (with the idea of ...Qf5 and ...Bxd6) - or even better is 20...a6! 21.Qb3 Kg7 and Black is fine. 21. Qb3?! (Fixated on this a2-g8 battery idea, I missed 21.Qd4! effectively winning, aiming to threaten Qxf7+ from f4 or f6, e.g. 21...Qa1 22.Ba2 Bg7 23.Qc4 Rf8 24.Re8 Qxb1+ 25.Bxb1 Rxe8 26.Kf1 Rc8 27.Qg4 etc.) - ...Bxd6 22.Bxf7+ Kg7 23. Be8
...Bc6 24. c4 Qd8? (24...Qa1 is more annoying for White, when there is no immediate way to force a favourable mass liquidation) 25. Qd1 winning. ...Qg5!? 26. bxc6 Qh5 a good practical try - I was glad I had ample time left to find the best (only) reply.
27. g3 Rxb1 28. Qxb1 Qxe2 29. cxd7 Bc5 30. Qa1+ Kh6 31. Qf6 Qd1+ 32. Kg2 Bb6
33. Qf8+ missing the simple 33.Bxg6! (...hxg6 gets mated quickly after 34.Qh8+) ...Kh5 34. h3 34.Qg7 was stronger. ...Qd4 35. g4+ Kg5 White to play and win!
In time trouble I couldn't see a win, so I could only shuffle about until the time control... 36. Qe7+ Kf4 37. Qf8+ Kg5 38. Qe7+ Kf4 39. Qf7+ Kg5 40. Qf3 Qxc4 and luckily, the same winning motif from move 36 (as well as 38, 40) is still available here.
41. Qe3+ Qf4 (41...Kh4 42.Qh6#; 41...Kf6 42.Qxb6+; 41...Bxe3 42.d8=Q+ Kh6 43.Qh4+ Kg7 44.Qe7+) 42. d8=Q+ 1-0 (42...Bxd8 43.h4+ Kxg4 44.Bd7+)