Annotated Game #70: Early endgame struggle

This next tournament game is from the first round of a quad.  My Class A opponent chose early to head for a queenless middlegame, which I think mostly benefited Black.  Some interesting tactical and positional themes arose at various points and in the final position I had the only winning chances, but accepted a draw due to the ratings difference (over 250 points).  While an understandable decision, this really isn't the way to improve one's chess, which requires the mental toughness to take on and defeat superior opponents.

Some highlights of the analysis:

• The White line with 4. a4 is considered a sideline of the Slav with 3. Nc3, apparently with good reason.  Black scores quite well in it and is not seriously challenged.  The counterblow 4...e5 is quite effective here.
• The tactic on move 9 that White missed is instructive.  The White knight can simply run roughshod over Black's queenside, which is undeveloped, with the dual threat of Nc7+ and Nb6.  The intermediate bishop capture on d2 for Black doesn't help.
• Black shied away from concrete analysis on move 18 of the obvious pawn advance, kicking the Nc3 and winning a pawn on d5 after the exchanges are through.  The actual move played, 18...Nd4, in fact invalidates Black's potential tactic by blocking the pin on the d-file.  This shows how my thinking on tactics was in the past much more muddled; I was unable to clearly break down the tactical elements in a position.
• Black keeps plugging away, however, and makes the good strategic choice to simplify down into a minor piece endgame where by move 27 his pieces are relatively stronger.
• The move 33 variation with ...h5 is an excellent example of endgame strategy and tactics.  Black could have assured his superiority on the kingside with this tactic.
• The move 39 variation has a game-winning tactic based on promotion and a unique X-ray motif.  Another useful pattern, along with the move 33 variation, to keep in mind for potential endgame tactics.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Class A"] [Black "ChessAdmin"] [Result "1/2-1/2"] [ECO "D10"] [Annotator "ChessAdmin/Fritz/Houdini"] [PlyCount "85"] {D10: Slav Defence: cxd5 (without early Nf3) and 3 Nc3} 1. d4 d5 2. c4 c6 3. Nc3 dxc4 4. a4 {White evidently doesn't want to play a gambit line and looks to recover the pawn as soon as possible.} e5 {end of personal book. The idea is of course similar to other lines in the Nc3 variation, for example after 4. e4, giving back the pawn in order to disrupt White's play.} 5. dxe5 Qxd1+ 6. Nxd1 Bb4+ {Black scores over 50% from this point with all moves in the database, but especially well with this one.} 7. Nc3 b5 $6 $146 (7... Nd7 { is the best by test, although there are very few games in the database. This was also Fritz's preference. Here's the highest-level example:} 8. Nf3 Nc5 9. Be3 Nb3 10. Rd1 Bf5 11. Nd4 Bg6 12. g4 h5 13. Bg2 Ne7 14. h3 hxg4 15. hxg4 Rxh1+ 16. Bxh1 Nc5 17. Ra1 O-O-O 18. Bg2 a5 19. f4 Be4 20. Kf2 Bxg2 21. Kxg2 Bxc3 22. bxc3 Nd5 23. Kf3 Rh8 24. Bf2 Rh3+ 25. Bg3 Nxc3 26. Kg2 Rxg3+ 27. Kxg3 Nb3 28. Nc2 Nxa1 29. Nxa1 Nxa4 30. f5 Nc5 31. Kf4 b5 {0-1 (31) Gavrilov,O (2338) -Danielian,E (2433) Minsk 2005}) 8. Bd2 Bd7 $6 {Black's intent when playing ... b5.} (8... Ne7 {would cover d5 and keep things relatively even.} 9. Nf3 $14) 9. e4 {This overlooks a tactical resource for White where he could exploit the d5 hole with his knight following the pawn exchange on b5.} (9. axb5 $5 cxb5 10. Nd5 Bxd2+ 11. Kxd2 {now Black cannot adequately defend against the knight's threats, as his rook is trapped in the corner.} Kd8 12. Nb6 axb6 13. Rxa8 $16) 9... a5 {now Black's rook has an out.} 10. axb5 cxb5 11. Nd5 Bxd2+ 12. Kxd2 Ra7 13. Ne2 (13. Nf3 {would give the knight greater scope and not block the Bf1.}) 13... Nc6 14. f4 Nge7 {a position of dynamic equality. } 15. h3 {Secures g4, notes Fritz. However, the bishop development to g2 takes some extra time, which Black uses to good effect.} (15. Nxe7 {is Houdini's preference, which would avoid later problems down the d-file.}) 15... O-O 16. g4 Rd8 17. Bg2 ({Bailing out with} 17. Nxe7+ {was still possible.}) 17... Be6 { putting White in a bind.} 18. Nec3 $2 (18. Kc1 {is what the engines agree is the best defense.} Rad7 (18... Nxd5 19. exd5 Bxd5 20. Bxd5 Rxd5 21. Nc3) 19. Nxe7+ Rxe7 $15) 18... Nd4 $6 {an overly cautious move that lets the Nd5 off the hook by breaking the pin.} (18... b4 {is the obvious and best response.} 19. Na4 Nxd5 20. exd5 Bxd5 21. Bxd5 Rxd5+ $17 {and Black is a clear pawn to the good with his pieces better placed than White's.}) 19. Rad1 $11 b4 { a move too late.} 20. Nxe7+ Rxe7 21. Nd5 {Now instead of being a pawn up, Black has another Nd5 to contend with.} Nc6 (21... Red7 $5 $11) 22. Kc2 Red7 { Black's pieces look menacing, with the pressure on the d-file and d5, but White has everything covered.} 23. Rhe1 b3+ {this is a bit impatient.} (23... a4 {is what Houdini advises.}) 24. Kb1 $6 (24. Kc3 {would take advantage of the new hole on c3.}) 24... Nb4 $15 {The black knight is well posted, comments Fritz.} 25. Nc3 Rxd1+ 26. Rxd1 Rxd1+ 27. Nxd1 Nd3 {without the rooks on the board, Black's well-placed knight is relatively stronger.} 28. f5 {interesting that it took White such a long time to play this. Despite the loss of the e5 pawn, White needs to extend his space on the kingside and drive Black's bishop away.} Bc8 29. Ne3 Nxe5 30. Bf1 Bb7 (30... Ba6 {is also possible, but I felt this would tie Black to the c-pawn in an uncomfortable way.}) 31. Nxc4 (31. Bxc4 $5 {would be the better way to capture, according to the engines.} Bxe4+ 32. Kc1 Nxc4 33. Nxc4 a4 $15) 31... Nxc4 32. Bxc4 Bxe4+ 33. Kc1 Bg2 (33... h5 { played immediately is a more sophisticated way to attack White's kingside pawns.} 34. Bxb3 h4 {the key move. Now the h-pawn will fall to Black's bishop and give White winning chances .}) 34. h4 Bh3 35. Be2 h5 36. g5 Bxf5 37. Bxh5 { In contrast with the above variation, here black's extra queenside pawn gets him very little, with proper defense.} Kf8 38. Bd1 a4 39. Kd2 $2 {a major tactical error, which I fail to spot.} (39. Be2 Bd7 $15 {and it looks quite drawish.}) 39... Ke7 (39... a3 $1 40. bxa3 (40. Bxb3 axb2 41. Bc2 b1=Q 42. Bxb1 Bxb1) 40... b2 41. Bc2 b1=Q {and the x-ray motif wins the bishop.} 42. Bxb1 Bxb1) 40. Kc3 Bd7 (40... f6 {immediately is a better try for an advantage, hoping to exchange off pawns and then pick one up with the king.}) 41. h5 f6 42. gxf6+ $2 (42. h6 $5 {is the best option White has, notes Fritz.} gxh6 43. gxh6 Kf8 44. Bh5 f5 45. Bg6 Kg8 $11) 42... Kxf6 $19 43. Kb4 {and here Black takes the draw due to the ratings difference, although he was in no danger and could have played on to see if there were winning chances.} 1/2-1/2