Annotated Game #108: Opening preparation?

This is the final game from the last Swiss tournament in the Slow Chess League.  One of the interesting features about playing in a Chess.com league is that all of your opponents' games on the site are accessible and downloadable.  This naturally can work both for you and against you.

It seemed to me at the time that my opponent must have looked at the previous round's game (Annotated Game #107) as part of his preparation.  The idea of playing the ...e4 advance as in the previous game could perhaps be improved by preventing the response Ng5 (which wins a pawn by force).  This appeared to be the idea behind Black's 4...h6 in this game, which otherwise has little point.  The opening takes a very different path from the previous game, but unfortunately I make some similar kinds of errors, including following a dubious and uncertain plan, which allows Black to take over the initiative and create too many threats for me to find my way through.

In general, I felt that my last two opponents were better focused during the games and wanted to win more than I did, so by that measure they certainly deserve their results.  I hope to do better on that score in my next games.

[Event "Slow Swiss #9"] [Site "Chess.com"] [Date "2013.11.03"] [Round "5"] [White "ChessAdmin_01"] [Black "Gunners2004"] [Result "0-1"] [ECO "A28"] [WhiteElo "1443"] [BlackElo "1718"] [Annotator "ChessAdmin/Houdini"] [PlyCount "52"] [EventDate "2013.??.??"] [TimeControl "45"] 1. c4 e5 2. Nc3 Nf6 3. Nf3 Nc6 {the English Four Knights} 4. e3 h6 {this is not as effective as Black's normal continuations that assist development, such as with ...Bb4 or ...Be7. The only benefit of the text move is to prevent a possible Ng5.} 5. d4 {the response of choice among master players in the database. White seizes the chance to strike in the center, where Black has less support than he should, after the previous move.} exd4 6. Nxd4 {this is the only move played in the database, but the pawn capture can be considered as well.} Bb4 7. Nxc6 {played with the idea of exchanging both the Nc6 and Bb4. } (7. Be2 {was played in the only victory (for White) in the small game sample. Houdini also likes it. White ignores the threat of doubled c-pawns, knowing that he will receive the two bishops in return, and gets on with development.} Bxc3+ 8. bxc3 O-O 9. O-O $11) 7... bxc6 8. Bd2 O-O 9. a3 Bxc3 10. Bxc3 $11 { White now has a pleasant game with the two bishops, although possesses no real advantage.} Ne4 {Black looks to force an exchange of White's excellent bishop on the long diagonal. Houdini spots a tactical, sacrificial alternative to simply allowing the exchange on c3, based on the fact that the Ne4 is hanging.} 11. Qd4 (11. Bxg7 $5 Nxf2 (11... Kxg7 12. Qg4+ Ng5 13. h4 d6 14. Qd4+ Kg8 15. hxg5 Qxg5 16. O-O-O $14) 12. Qf3 Kxg7 13. Qxf2) 11... Nxc3 12. Qxc3 f5 13. g3 { played to control f4, otherwise the Black f-pawn can be used as a battering ram. Part of the plan for White is castling queenside.} Rb8 14. Bd3 (14. c5 { is something I had considered, as it fixes Black's c-pawn and opens up the a2-g8 diagonal for the bishop, which is less effective in the game continuation. I also was stubborn about sticking to the idea of queenside castling, to the detriment of other positional factors.}) 14... c5 15. O-O-O $6 $15 (15. O-O {and White is safe enough, despite the light-square weakness, since it is difficult to exploit. White could then have pursued a queenside expansion strategy, for example} Bb7 16. b4 Qe7 17. Rab1 $11) 15... Bb7 { Black now effectively takes over the initiative.} 16. Rhg1 Bf3 17. Rd2 Rf6 18. Be2 Bxe2 19. Rxe2 Qe7 20. Rge1 {I had a long think here, as White does not have an obvious plan and multiple weaknesses to cover.} (20. Rd1 {was another alternative considered and probably simpler and better. At the time, the hanging Re2 bothered me and I did not see a good follow-up for White. Play could continue} d6 21. Qd3 {with a number of possibilites, but White should be OK.}) 20... Rfb6 (20... Qe4 $5) 21. e4 $11 {this felt risky at the time, but I calculated that it would work. I was correct, but unfortunately went astray afterwards.} Rb3 {I had in fact missed this move, focusing on the variations with the exchange on e4, and failed to find the correct continuation.} 22. Qa5 $6 (22. Qd2 {I wrongly rejected, fearing a possible breakthrough against White's king.} fxe4 23. Rxe4 Qf7 24. Qf4 Qxf4+ 25. Rxf4 Rxb2 26. Re7 Rb1+ 27. Kc2 {and Black has nothing more than a perpetual check with the two rooks.}) 22... fxe4 23. Qxc7 (23. Rxe4 {might have provided better practical chances.} Qf6 24. Re8+ Kh7 25. Qd2 Rxb2 26. Qd3+ Qg6 27. Qxg6+ Kxg6 28. Rxb8 Rxb8 $15 { and White has a pawn-down rook ending to defend.}) 23... Qf6 24. Qxd7 $2 { this leads to an immediate collapse, but I didn't see anything better at the time.} (24. Qf4 {would make things much more difficult for Black.} Rc3+ (24... Rxb2 {no longer works due to} 25. Qxf6) 25. Kb1 Qxf4 26. gxf4 Rxa3 27. Rxe4 $15 ) 24... Rxa3 25. Qd5+ Kh8 26. Rc2 {White is lost at this point.} (26. bxa3 Qc3+ 27. Rc2 {and I had previously missed in my calculations that the Qc3 could now capture on e1.}) 26... Ra1+ 0-1