IM K. Regan v NM L. Davis

IM Ken Regan was kind enough to share his own annotation of his recent game with NM Lionel Davis from the recent KLCC January Tournament in Buffalo. He noted that "the game has 4 nice instructive moments, 2 each where [either of us] could have gained a strong or winning advantage. [It] also shows [the] differences of opinion among 3 chess programs!" 1. e4 d5 2. exd5 Qxd5 3. Nc3 Qd6 4. d4 Nf6 5. Nf3 Bg4 6. Bc4 {The "book move" is 6.h3 to disturb the Bishop right away. There's nothing wrong with White's developing moves, but they let Black "get comfortable."} c6 7. O-O Nbd7 8. Be3 e6 9. h3 Bh5 10. Qe2 Be7 11. Rfe1 O-O 12. Rad1 Rfe8 13. Bb3 Qc7 {Fritz 9 and HIARCS 11.2 both definitely like Black, but several Rybka versions give 0.05-0.10 to White, liking my enterprising next move. At the board I thought I had a chance for an edge, and played confidently.} 14. g4 Bg6 15. Ne5 Nxe5 16. Bf4 {White will regain the piece, but Black has many choices: 16...Nd5, 16...Nfd7 and 16...Bb4!? which could be annoying. Lionel played the move I expected, not letting me get my pawn to e5.} Nf3+ 17. Qxf3 Bd6 18. Be5 Nd5 19. Qg3 Bxe5 20.dxe5 Nxc3 21. Qxc3 {White has occupied e5, but Black has chopped off the pieces that were supposed to come in behind the pawn. All the engines now say the game is pretty much equal, but I forged ahead with my plan, which was now to use e5 to invade on d6. This plan has a tactical flaw which Black only exploited halfway...} Rad8 22. Qg3 c5! 23. Rd6? {The first instructive moment. Black to play and punish White's plan.} {The flaw was 23... c4 24. Ba4 (24. Rxd8 Rxd8 25. Qc3 Rc8 26. Ba4 a6 {is also lost for White}) (24. Qc3 b5 25. a4 Rxd6 26. exd6 Qxd6 27. Ba2 { is how White has to play, but Black is a Pawn up with a better position too.}) 24... b5 (24... Qa5 {also forks but falls into White's plans} 25. Red1 Rxd6 26. exd6 Qxa4 (26... Rd8 27. d7 {is unclear---Fritz gives Black a pawn edge but with a line I think is really risky. Fritz 9:} b5 28. Qe5 a6 29. Qc5 h6 30. b4 Qxa4 31. Qc8 Rxc8 32. dxc8=Q+ Kh7 33. Rd8 {now Fritz sees what I meant by "unclear"} 28. Qc7 {wins, because Black's Queen can't get to b6}) 25. Bxb5 Qa5 {and now after White cedes the piece to make a strong d-pawn, Black's Queen will be more active than on a4.} 26. Red1 Rxd6 27. exd6 Qxb5 28. d7 Rd8 29. Qc7 Qb6 30. Qxc4 Qb7 {is the point, covering everything and preparing ...Be4.}) 23...Rxd6? 24.exd6 Qc6 25. Qf4 Rd8 26. Rd1 {Now Fritz 9 agrees with Lionel's after-game belief that White was still overextended, saying sizable edge to Black, but the Rybkas give White the sunny side of equality, while HIARCS is neutral. Maybe the engines should be voting in next week's primaries.} b5 27. c4 {Black may be able to equalize by 27...bxc4 28.Bxc4 Be4, but with time getting short such a position feels "loose". The move actually played gives White a big tempo, however.} b4 {The second instructive moment. Now I saw Black's idea of ...f6 and ...e5 to orphan White's d-pawn, and with about 4 minutes left, figured the best reaction naturally had to be improving my Bishop onto the long diagonal. But White needs to use the opportunity which Black's slow last move gave!} 28. Rd2 (28. d7 f6 (28... h6 29. Rd6 (29. Qd6 Qxd6 30. Rxd6 {is also a slight pull for White, but not as strong as after 28...f6 weakening e6.}) 29... Qc7 30. Ba4 {and now White's Bishop is a smooth operator.} Qa5 31. b3 $16) (28... Rxd7 29. Qb8+ {is mate}) 29. Qd6 $1 (29. Rd6 Qc7 30. Ba4 Bd3 $1 31. Bc6 $1 h6 32. Rxe6 $1 Qxf4 33. Re8+ Kh7 34. Rxd8 Qc1+ 35. Kh2 Qc2 36. Re8 Qxf2+ $11) 29... Qxd6 30. Rxd6 Kf7 31. Ba4 Ke7 32. Ra6 Bd3 33. Bb5 $16 { Black's real problem is that all 3 of eir Q-side pawns are weak.}) 28... f6 29.Bd1 e5 30. Qf3 (30. Qe3 Rxd6 31. Bf3 Qb6 {gives activity but not enough for the Pawn. My idea was always 30.Qf3 even though it seems to lose a tempo---I just didn't see the full followup.}) 30... Be4 (30... Rxd6 {may be what Black should play, but I disagree with engines and like the following line for White: } 31. Rxd6 Qxd6 32. Qa8+ Qf8 33. Qxa7 Bd3 34. Qa6 Kh8 35. a4 bxa3 36. bxa3 h6 37. Bf3 e4 38. Bg2) 31. Qe3 {It looks like Black has taken over the main diagonal, but this is an illusion---the engines all agree by "Depth=15" or so that Black must now meekly give it back by 31...Bg6.} 32. Rxd6? {The third instructive moment, and the easiest one to answer. White to move and win!} Beter instead was 32. Ba4! and now Black cannot defend both Rook and Bishop! The engines think Black is semi-OK giving the Queen for Rook and Bishop, but Black's Queenside pawns are too weak to hold on. Rxd2 33. Bxc6 Rd1+ 34. Kh2 Bxc6 35. f4! (35. Qxc5? Rh1+ 36. Kg3 Rg1+ {is a perpetual since} 37. Kh4 g5+ 38. Kh5 Rh1! 39. Qe3 Rxh3+ 40. Qxh3 Kg7 {is an unbelievable forced mate!}) 35... Rh1+ 36. Kg3 Rf1 {and now comes an ending that I would certainly have demonstrated given enough remaining clocktime...} 37. fxe5 $1 (37. Kh4{looks risky but is fine too}) 37... Rf3+ 38.Qxf3 Bxf3 39. Kxf3 Kf7 40. Ke4 Ke6 41. exf6 gxf6 {Study this, and convince your self before your engine realizes that this is winning for White. White will trade off the K-side so that his King will end up on g5, which will give clear sailing over to the west to win Black's c-pawn. Black may be able to preserve the h-pawns and so win White's h-pawn and queen eir h-pawn right after White queens eir c-pawn, but the Queen ending will be hopeless as the remaining Black pawns over there fall.}) 32... Qxd6 33. Qxe4 Qxd1+ 34. Kg2 Qd7{I feel that this should still not be losing for White, given the active White Queen, but Queen endings a pawn up are winning more often than I used to think.} 35. Qa8+ Kf7 36. Qe4 g6 37. b3 Ke6 {Brave, maybe not right, but Black's King has the extra benefit that coming forward thru White's weak squares might both hide from checks and assist in a mating attack, as could and should have happened.} 38. Qa8 f5 {Ditto. But it challenges White...} 39. g5? ({I didn't see that} 39. Qg8+ {was already enough activity to win back a pawn, when White actually stands better!} Qf7 (39... Kf6 40. Qf8+) (39... Kd6 40. Qd5+) 40. Qc8+ Kd6 $2 41. Qa6+ {oops!}) 39... e4 40. h4 Ke5 41. Qb8+ Qd6 (41...Kd4 42. Qf4 $1 Qc6 {preserves Black's edge}) 42. Qxa7 f4 {Now Black can think of mate, and I made an instructive mistake, giving Black's King free access to the g4 square. This is very similar to a losing error at move 35 by Boris Gelfand against Tejmour Radjabov in round 11 of the just-finished Wijk-aan-Zee tournament, as explained here: http://www.chessbase.com/newsdetail.asp?newsid=4413} 43. Qg7+? (43. Qf7 {keeps Black at bay---engines show basic equality after ...Qc6 or ...f3+.})43... Kf5 44. Qxh7 Kg4! {Now trouble! At least I saw it wasn't instant mate.} 45. Qb7 f3+ 46. Kh1 {The fourth instructive moment. "Always check, it may be mate" is not a law of chess. But with Black down to just the 5 seconds per move given by the Fischer time increment, a quiet winner is tough to spot...} 46...Qd1+ (better was 46... Qd4! 47. Qc8+ Kxh4 $19 {The Queen on d4 guards the check at h8,and after White's next comes another quiet move for the coup-de-grace.} 48. Kg1 e3! {Mate in 6, the engines say.}) 47. Kh2 {Alas, now after 47...Qd4 48.Qc8+ Kxh4, White has 49.Qh3+ and maybe salvation. Otherwise, Black has nothing better than the perpetual check on d1 and d6 which fairly reflected the chances both players had in the game.} 1/2-1/2