GOTW: FM Dougherty vs. GM Marin

[Event "Reykjavik Open"]
[Site "Reykjavik, Iceland"]
[Date "2016.03.09"]
[Round "3.22"]
[White "Dougherty, Michael"]
[Black "Marin, Mihail"]
[Result "0-1"]
[ECO "E94"]
[WhiteElo "2169"]
[BlackElo "2597"]
[Annotator "MacKinnon,Keith"]
[PlyCount "114"]
[EventDate "2016.??.??"]

1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 {committing to the King's Indian Defense. It's a
good opening for when you want to play for a win (often at all cost)} 4. e4 d6
5. Nf3 O-O 6. Be2 e5 7. O-O Na6 $5 {I give the interesting annotation because
this is the first time that GM Marin has played the move (at least according
to Mega 2015). An experienced professional, Marin knows that in playing lower
rated opponents, it is important to avoid long prepared lines. He has essayed
the more common 7...Nc6 and 7...Nbd7 in the past} 8. Be3 c6 (8... Ng4 {likely
the most testing variation, but I suspect Marin did not want to enter forcing
lines - as would occur after} 9. Bg5 Qe8 10. dxe5 dxe5 11. h3 h6 (11... f6)) 9.
d5 (9. Re1 {another reasonable option which is less committal than d5} Ng4 10.
Bg5 Qe8 11. h3 f6 12. Bc1 Nh6 13. Bf1 {and we reach a strange situation where
both White bishops return to their starting squares, yet White is in the
driver's seat here and scores about 65%}) 9... Ng4 10. Bg5 f6 11. Bh4 c5 12.
Nd2 (12. Ne1 {is played more often and would likely be my preference; however,
the following game shows that here, too, Black can get a strong position} h5
13. a3 Nh6 14. h3 Nf7 15. Nd3 Bh6 16. b4 Bd7 17. Rb1 b6 18. bxc5 Nxc5 19. Nxc5
dxc5 20. a4 Nd6 21. Qd3 Kg7 22. Rb2 Bg5 23. Bg3 h4 24. Bh2 Qe8 {0-1 (65) 0-1
Wojtaszek,R (2744)-Ding,L (2732) Wijk aan Zee 2015}) 12... Nh6 (12... h5 {
looks more natural to me and has been played a number of times. I'm not a big
KID expert though and I'm sure Marin had his reasons for the Knight retreat})
13. f3 {Probably the first move I disagree with. I understand FM Dougherty's
desire to support e4 and have an escape route for his Bishop on h4, but this
just seems somewhat passive to me} (13. a3 Nf7 14. Rb1 {would probably be how
I would like to continue here. As is typical for the KID, White's plan
revolves around putting pressure on Black's queenside while parrying an attack
on the kingside.}) 13... Nf7 14. Bd3 $146 ({The CCN editor, John Upper, points
out the following crushing win for Black by GM Shabalov back in 2010} 14. a3
Bd7 15. Rb1 Bh6 16. Bf2 f5 17. b4 b6 18. Nb5 Bf4 19. bxc5 Nxc5 20. Nb3 fxe4 21.
Nxc5 dxc5 22. fxe4 Bxb5 23. cxb5 Nd6 24. Bf3 Nc4 25. Ra1 Nd2 26. Re1 c4 27. Ra2
c3 28. Rc2 Rc8 29. Qe2 Qg5 30. Bg4 Rc4 31. h4 Qf6 32. Qd3 $2 Bh2+ $1 $19 33.
Kxh2 Qxf2 34. Qg3 Qxg3+ 35. Kxg3 Rxe4 36. Rec1 Nb3 37. Bf3 Rd4 38. Re1 Rd3 {
0-1 (38) Shankland,S (2507)-Shabalov,A (2585) Philadelphia 2010}) 14... Bh6 15.
Qe2 Kg7 16. a3 Qe8 {The Black queen moves off the h4-d8 diagonal in order to
play an ...f5 break} 17. g4 $2 {this just seems like a major strategic error
and overreaction. It was not so serious to allow Black to get f5 in. White
should have continued with the aforementioned plan of attemping to play b4} (
17. Bf2) 17... Bf4 $1 {White has created dark-squared weaknesses in his camp
and has to contend with ideas of ...Rh8 and ...h5 coming very shortly} 18. Kh1
Rh8 (18... h5 $1 19. gxh5 Rh8 20. Rg1 Rxh5 21. Bg3 Bd7 $15 {and Black is
firmly in the driver's seat. He will aim to play ...Qe7 and ...Rah8} ({note
that if Black goes with the natural move, ...Ng5, bringing an extra attacker
into the position, White has an unbelievable resource (which I cannot take
credit for - Komodo found it in about 5 seconds)} 21... Ng5 $2 22. Bxf4 exf4
23. e5 $3 dxe5 24. Nde4 $14 {the pawn sacrifice in order to open up the d3
bishop and put the Knight on the very powerful e4 square offers more than
enough compensation. Black's king position now begins to feel slightly suspect}
)) 19. Rg1 h5 20. Bg3 (20. gxh5 Rxh5 21. Bg3 {would transpose to the variation
above}) 20... Rh6 $6 {I believe that both players made a number of
inaccuracies starting around here. They both seemed reluctant to do anything
about the tension on g4} (20... hxg4 21. fxg4 Bxd2 22. Qxd2 Bxg4 {is very
tempting, but I guess Marin preferred to keep the pressure up. The loss of the
g4 pawn does ease White's task to some extent, however I don't see very much
compensation for White} 23. Raf1 Ng5 24. Be2 Bxe2 25. Qxe2 Nh3 $17) 21. Nf1 $6
(21. gxh5 $1 {While this move does open up the h-file, that will happen
anyways, and here there is a tactical justification} Rxh5 $2 (21... Bxg3 22.
Rxg3 Rxh5 23. Rag1 Ng5 24. Qe3 {And, although Black has some pressure, the
position should be fairly equal. The computer even recommends Black repeat
moves after} Qh8 25. R1g2 Bh3 26. Rg1 Bd7 27. R1g2 Bh3) 22. Bxf4 exf4 23. e5 {
I wonder if White didn't see this resource. Here, White is almost winning} f5
$8 24. e6 Ne5 25. Nb5 $18) 21... Bd7 (21... hxg4 22. fxg4 Ng5 $40 {looks very
strong}) 22. Ne3 Qe7 (22... Bxe3 23. Qxe3 hxg4 24. fxg4 Bxg4 {looks like a
clear advantage to Black. There may be nuances to the position that I don't
understand, but I really don't know why Marin didn't take the material}) 23.
gxh5 $1 g5 $6 {Surely, this is taking it too far. Now White has the option of
plopping a Knight on e4 after Nf5+} 24. Ng4 {A strange place for the Knight} (
24. Nf5+ Bxf5 25. exf5 Qd8 {The Queen needs to get off the e-file as the
Bishop would otherwise be hanging on f4 due to the two pins}) 24... Rxh5 25.
Bxf4 $2 (25. Nd1 {heading for e3 could have been an idea}) 25... exf4 26. e5 $5
{Not as effective as in some of the lines previously, but probably still a
decent try to mix up the position} Nxe5 (26... dxe5 {may have been more
difficult to handle. A Black Knight landing on d6 would be tough to deal with 
(As a general rule, Knights do well blockading passed pawns)}) 27. Nxe5 dxe5
28. d6 {Wow. Double pawn sacrifice to get some play - it's bold, but I doubt
it's completely sound} Qf7 $6 {Marin probably wanted to win cleanly, but I
suspect that taking on d6 was safe enough. A sample line is as follows:} (28...
Qxd6 29. Rad1 Qe7 30. Be4 Rah8 $19) 29. Be4 Be6 (29... Rah8 30. Bxb7 Rxh2+ 31.
Qxh2 Rxh2+ 32. Kxh2 Qh5+ 33. Kg2 Qh4 34. Kf1 Nb8 $19) 30. b3 Rah8 31. Rg2 Nb8
$1 {A strong prophylactic move against Nb5, and it prepares to improve the
Knight's position} 32. Rd1 b6 (32... Qd7 {with the idea of ...Nc6 and Nd4
looks good as well}) 33. Nb5 Qd7 34. Qd3 Bh3 35. Re2 {At first blush, it might
be tricky to see that this is a mistake, but White needed to act swiftly with
Rb2 and b4} (35. Rb2 Nc6 36. b4) 35... a6 $6 {A bit of an inaccuracy since now
the White Knight gets to the d5 square} (35... Nc6 $1 36. Rb2 (36. Bxc6 Qxc6
37. d7 Rd8 $19) 36... Nd4 37. b4 f5 38. Bd5 Nxb5 39. cxb5 Qxd6) 36. Nc7 {White
starts playing quite well here} Nc6 37. Bxc6 Qxc6 38. Nd5 R5h6 (38... Qxd6 $2 {
would be a good way to throw the win away!} 39. Nxf4 Qxd3 40. Nxh5+ Rxh5 41.
Rxd3 $14 {and now Black fights (although not too hard) for the draw}) 39. Ne7
Qd7 40. Rg1 Qe6 {The dreaded 40th move} (40... Be6 {was preferable}) 41. b4 Rd8
42. Rd2 Rd7 43. Rgd1 Kf7 44. bxc5 bxc5 45. Qe4 f5 46. Qa8 Kg7 47. Rd5 $4 {The
losing mistake. White would have liked to put pressure on e5 with Re1 instead.}
Rh8 48. Qc6 Rxe7 49. Qxc5 Rd7 (49... Rc8 $2 {would have bungled it!} 50. Qxc8
Qxc8 51. dxe7 g4 52. Kg1 $1 {and, somehow, this is a draw, as White is
threatening Re5 or e8=Q and then checks with the rooks} (52. e8=Q Qxe8 53. Rd7+
Kf8 54. Rd8 gxf3 $19 {would be a painful way to go down as White})) 50. Rxe5
Qh6 51. Re2 Rhd8 52. Qe5+ Qf6 53. c5 Qxe5 54. Rxe5 Kf6 55. Rde1 g4 56. fxg4
Bxg4 57. h4 $2 (57. Re6+ Kf7 58. Kg2 {and I'm not completely sure the Black
position is won. However, White's margin for error would certainly be very slim
}) 57... Bf3+ {What a topsy-turvy game!} (57... Bf3+ {FM Dougherty didn't need
to see} 58. Kg1 Be4) 0-1