In accordance with the long-standing Gibraltar tradition, the select few still fighting for the 1st prize start their final games at 10am to accommodate possible tiebreaks, which, combined with the pairings announced after 10pm, creates a toss-up choice between proper preparation and sleep. I opted for freshness - Hari plays a wide variety of openings these days.
1.e4
e5
2.Nf3
Nc6
3.Bb5
a6
4.Ba4
Nf6
5.0-0
Be7
6.Re1
b5
7.Bb3
d6
8.c3
0-0
9.h3
Nb8
The Breyer was almost the only thing I looked at before the game - it seemed like Hari's weapon of choice in particularly important games. I've also played it a few times myself recently.
10.d4
Nbd7
11.Nbd2
Bb7
12.Bc2
Re8
13.a4
However, this was an over-the-board decision. My morning was spent on trying to make sense of the recent developments in the [13.Nf1
Bf8
14.Ng3
g6
15.b3
line - which we both faced with Black against Karjakin recently, and had very few problems in the opening. Two rounds previously this position was revisited in the game between Hikaru Nakamura and Vitaly Golod, and I spent a few moments trying to figure out whether it made sense to repeat it. After 15...Bg7
16.d5
Rc8
Hikaru went for the old sharp line: 17.Be3
(Karjakin played (17.Bd2
c6
18.c4
against Hari and; 17.Bd3
c6
18.c4
against me) 17...c6
18.c4
Nb6
19.Qe2
cxd5
20.cxd5
Nbxd5
21.exd5
Nxd5
, and in this position, which is considered to be good for Black since Ju.Polgar-Spassky 1993, produced a novelty: 22.Rac1!?
Nc3
23.Qd2
Nxa2
24.Ra1
Nc3
25.b4
d5
26.Bb3
, getting his pieces ready to meet the pawn advance in the center. However, questions remained: apart from the ambitious 26...Qe7!?
chosen by Golod, Black could also play the simpler (26...e4
27.Nd4
Qd6
28.Nge2
Qxb4
, and although the avalanche is stopped, Black has 4 pawns to show for the piece. The game continued) 27.Bh6
Bh8
(27...Qxb4
would have been a mistake - 28.Bxg7
Kxg7
29.Nxe5!
, and the bishop cannot be touched: 29...Qxb3
30.Nh5+!
with forced mate to follow) 28.Nxe5!?
(a very committal decision, but white did not really have much choice) 28...Bxe5
29.f4
, and here Black went badly wrong: after 29...Ne4? 30.Rxe4! dxe4 31.fxe5 Qxe5 32.Rf1 Re7 33.Rxf7! Rxf7 34.Qd7 White was completely winning 29...Qh4
was necessary, and after 30.fxe5
Qxg3
31.Qd4!
(31.Re3
Qh4
is quite bad for White - he needs to control the h4 square before playing Re3) 31...Re6!?
seems to be best, since the natural (31...Rc4
runs into 32.Qb6!?
(32.Bxc4
dxc4
33.Qd2
is also possible, but Black has an immediate draw after 33...Ne4
34.Qd7
Qf2+
35.Kh2
Qg3+
and can look for more - perhaps after 33...Nd5) 32...Qh4
(32...Rh4?
33.e6!
loses on the spot) 33.Bxc4
(33.e6
seems to be a draw by repetition: 33...Qd4+
34.Qxd4
Rxd4
35.Rac1
Rd3
36.Bc2
Rd4
, and there is nothing better than 37.Bb3 Rd3 38.Bc2 ad nauseam) 33...dxc4
34.Qxb7
Qxh6
35.e6
, and White is better. After the forced) 32.Re3
Ne2+
33.Rxe2
Qxb3
34.Qa7!
Black will need to find 34...Qc3!
, since (34...Re7
35.Qb6
Bc6
36.Rc1
Re6
37.Qxa6
Rc7
38.Rec2
g5
39.Qb6!
is very good for White - but in the endgame arising after) 35.Rc1
Qxc1+
36.Bxc1
Rxc1+
37.Kh2
Bc6
Black is doing very well. I was sure Hari was aware of this game as well, and although the sharp nature of the position would have suited me just fine, I did not have enough faith in my last-minute analysis. In the end, I chose a much quieter line with which Hari had somewhat less experience.]
13...Bf8
14.Bd3
c6
15.b3
g6
16.Qc2
Rc8
17.Ba3
White normally develops this bishop to b2, but I always felt that this line is so balanced that it makes precious little difference which moves one makes, and the text move poses some fresh questions.
17...Qb6
18.Bf1
Continuing the re-grouping - and also defending against the thematic push in the center.
18...Nh5
[The tactical point of 18.Bf1 is seen in the following line: 18...d5?
19.Bxf8
Kxf8
20.exd5
exd4
21.Rxe8+
Rxe8
22.dxc6
Qxc6
23.Nxd4
, and the pawn on g2 is protected.]
19.g3
exd4
Perhaps bored by the waiting game played thus far, Hari forces the issue in the center. Another possible approach was [19...Ng7
20.Qd3
Ne6
21.h4
, and White exerts some pressure - the second bishop is about to enter the fray via h3.]
20.cxd4
c5
21.axb5
I spent quite some time here trying to figure out whether it is better to play Qd3 immediately, or open the a-file first. [After 21.Qd3
Black does not have the option of protecting the b5 pawn - but I decided that his best option in any case is 21...b4!?
22.Bb2
cxd4
23.Bxd4
Qc7
, starting immediate play against the e4 pawn - Black is ready to play Nc5 or Nhf6 next - and that I need to have the a-file open to have any chance of fighting for advantage.]
21...axb5
22.Qd3
Ba6
Not a bad choice in itself - but even in this position I was more worried about [22...b4!?
23.Bb2
cxd4
24.Bxd4
(24.Nc4
Qd8
25.Bxd4
(25.Qxd4?
Bg7
is bad: 26.Qxd6
Bxb2
27.Nxb2
Ne5!
) 25...Nc5
26.Bxc5
Rxc5
27.Bg2
Nf6
, and Black is perfectly fine) 24...Nc5
25.Qb5
(25.Qc4
Bg7
is also rather unimpressive) 25...Qxb5
26.Bxb5
Re7
, and Black has excellent chances of weathering the storm. The text move preserves Black's structure - but Black will have to constantly worry about bringing this bishop back into play.]
23.Bb2
d5!?
A very interesting attempt to equalize immediately - and understandable, too, since Black can ill-afford to wait: after [23...Bg7
24.Ra2!
, protecting the bishop and preparing to double along the a-file, Black is stuggling to coordinate his pieces. Black could try the same idea with a different move order:; 23...c4!?
24.bxc4
(24.Qc2
d5!
is less clear: 25.bxc4
(25.exd5
Rxe1
26.Nxe1
(26.Rxe1?
c3!
27.Bxc3
b4
) 26...Nhf6
, and Black is fine) 25...dxc4
(25...bxc4
26.Bc3!
) 26.Reb1
, and while I still like White, the position becomes very sharp) 24...bxc4
25.Qc2
d5
, but after the calm 26.Bc3!
White still retains some pressure: 26...dxe4
27.Nxe4
Bb7
28.Rab1
Qc7
29.Ne5
Bd5
30.Rb5
, and Black will have to be careful.]
24.dxc5!
I played this move quite quickly, believing it gives me a big advantage, and deciding not to waste too much time on [24.exd5
Rxe1
(worse is 24...c4
25.bxc4
bxc4
because of 26.Qb1!
(26.Qc2
Rxe1
27.Rxe1
c3
28.Bxa6
Qxa6
29.Bxc3
) 26...Rxe1
27.Nxe1
Nhf6
28.Qa2!
, and the pawn on c4 will fall) 25.Nxe1
Nhf6
, which would allow Black too much freedom]
24...Nxc5
[After 24...Bxc5
25.Bd4!
Black will struggle to equalize despite all the pawn trades in the center - the b5 pawn will be very weak in any resulting endgame]
25.Qd4
Bg7
26.e5
Qb7!
The only move - Black needs to be able to meet 27.g4 with Ne6. To my great surprise, the more I thought here, the less promising my position felt. Black will need to spend a bit of time to bring the h5 knight back into play - but if and when he does, his position will become quite playable. After 20 minutes of fruitless deliberation (and I would badly need this time later), I went with my original idea. In fact, my next two moves were both seriously inaccurate, and as a result most of White's advantage would be gone
27.Rec1?!
I rejected [27.Qe3!
in view of 27...Nf6
(27...Ne6
28.Nd4
Nxd4
29.Bxd4
favours White: 29...Bf8
30.Bg2
Ng7
31.Qd3
Ne6
32.Bb2
) , and Black succeeds in solving the wayward knight problem right away, but in fact White has an important resource here: 28.b4!
, and Black will be forced to put this knight on a4 since the e-file cannot be occupied while the other one is hanging: 28...Na4
29.Bd4
, and White is better.]
27...Ne6
28.Rxc8?!
Once again it was better to start with [28.Qe3
- I did not want to allow the trade of the horrible a6 bishop, but giving up on the c-file turned out to be no less of a concession. After 28...Rxc1
(28...Bf8
29.Nd4
Nxd4
30.Qxd4
Ng7
31.Bg2
does not promise Black much either) 29.Rxc1
b4
30.Bxa6
Qxa6
31.Nd4
Qa2
32.Rb1
the problems with the knight on the rim remain hard to solve]
28...Rxc8
29.Qe3
Bf8
30.Nd4
Less than ideal - but postponing it did not seem right either. After the possible [30.Bd3
Nhg7
Black is planning to play 31...Rc6 and 32...b4, while White will struggle to find a role for the d2 knight.]
30...Nxd4!
Of course, White can't be allowed to connect the knights.
31.Bxd4
Ng7
32.Nf3
Such was my desperation to save the bishop that I even considered playing [32.Ba7
here, but luckily for me the refutation of this ridiculous idea is not that hard to find: 32...Ra8!
(32...Nf5
33.Qb6
is the point) 33.Bc5?
Nf5
34.Qc3
b4
, and White is suddenly in trouble, but; 32.Bd3
deserved attention: 32...Ne6
33.Nf3
Nxd4
34.Qxd4
Bc5
35.Qf4
, and the bishop is still on a6, while White has some tangible initiative on the kingside. By this point I had a distinct feeling that I've misplayed a very nice position, and to compound my disappointment, I only had about 15 minutes left.]
32...Ne6?!
This move gave me renewed hope. I could not quite understand why Black would decline to trade the bishop with tempo, but I was not about to complain. [32...Nf5
was better: 33.Qd3
Nxd4
34.Qxd4
(34.Nxd4
Qb6
is completely harmless) 34...Bc5
35.Qd2
Qb6
36.b4
Bf8
, and Black is hanging on. White is still better after 37.Ra5
, but if Black could manage to shed one or even both of his weak pawns and activate his dormant bishop, he should be able to escape with a draw.]
33.h4
I did not waste too much time calculating this move - it will come in useful in most situations. Additionallly, White can now consider playing Bh3 and trading the light-squared bishop for the knight.
33...Nxd4
A decent move connected with the wrong idea. After the calmer [33...Qc6
34.Bh3
White is still on top: 34...h6
(34...b4
35.Ng5
Bh6
(35...Be7
36.Nxe6
fxe6
37.h5
, and White's advantage is growing fast) 36.Qd2
, and Black is having a hard time protecting his numerous weaknesses) 35.Bb2
Rc7
36.Rc1
Qb7
37.Rxc7
Qxc7
38.Nd4
, and Black will have to work hard to survive.]
34.Nxd4
Bc5?
This is pure recklessness. Black had to play [34...Qb6!
, where 35.e6
(or, perhaps, 35.Bh3!?
Bc5
36.Rd1
Re8
37.b4!
, and White is better due to the fact that 37...Bxb4
38.Bd7!
is very dangerous) 35...Bg7!
36.exf7+
Kxf7
37.Rd1
gives White a stable but hardly spectacular plus. (37.Qf3+
Kg8
38.Qxd5+?
Kh8
only plays into Black's hands) ]
35.b4!?
Once again, a decision that came pretty easily: a free hand in attack is much more important than a mere pawn. However, no-frills chess would also work: [35.e6
Qb6
(35...Re8
36.Qc3
Bxd4
37.exf7+
Qxf7
38.Qxd4
, and Black is in grave danger.) 36.exf7+
Kxf7
37.Qf3+
Kg7
, and now the move which, I suspect, we both missed: 38.Re1!
gives White a huge advantage.]
35...Bxb4
36.e6
Re8
The fact that, according to the machine, the only other move that does not lose on the spot here is 36...Ra8, should tell the reader a bit about the dangers of Black's situation.
37.Qb3!?
I was very proud of this geometrical motiff - and it does win. However, it was much cleaner to start with [37.Bg2!
, and now: 37...f6
(37...Bc5
loses on the spot: 38.exf7+
Qxf7
(38...Kxf7
39.Qf3+
Kg8
40.Nb3+-
) 39.Bxd5
Rxe3
40.Bxf7+
Kxf7
41.fxe3+-
; 37...fxe6
38.Qb3
transposes to 37.Qb3) 38.e7!
(you will see this idea later - but here it is much easier to find the follow-up) 38...Bc5
(38...Rxe7
39.Qb3
) 39.Qe6+
Kg7
40.Nb3
, and Black loses material]
37...Bc5
38.Bg2
f6!
An excellent practical choice. Judging by the heartless decimals of the computer's evaluation, Black had more chances to survive after [38...fxe6
39.Nxe6
Bxf2+
40.Kh2!
(but not 40.Kh1?
Qc8!
41.Bxd5
Bb7
) 40...Bxg3+
41.Kxg3
Qb8+
42.Nf4
Bb7
43.Bxd5+
Kg7
, and there is still some work to be done. I did see all of this at the board, though - none of the decisions White has to make in this line are too taxing. The text move loses by force - but with less than 5 minutes left at this point, I flunked that particular test.; 38...Rd8
, meanwhile, loses to 39.exf7+
Qxf7
40.Nc6!
Qxf2+
41.Kh2
, and once again Black begins to lose bishops.]
39.Qc3?
Played with seconds remaining on the clock. I knew this was probably not enough for the win, but I simply could not find anything after [39.e7
Kh8
The solution turned out to be quite spectacular: 40.Ne6
Bd6!
(I did not even see that 40...Bxe7
loses to 41.Qa2!
, to be brutally honest) 41.Nd8!
Qxe7
42.Nc6
Qb7
43.Qxd5
Be5
(43...Bf8
44.Qa2
is a recurring motiff) 44.Nxe5
Qxd5
45.Bxd5
Rxe5
, and now 46.Bc6!
, and with four pieces remaining on the board White wins the bishop yet again]
39...Be7??
Inexplicable. With about half an hour left on the clock, Hari played this move almost instantly, when not one but two alternatives would most likely save the game for him: [39...Qb6
40.Rxa6
Bxd4
41.Rxb6
Bxc3
42.Bxd5
Re7
43.Rxb5
f5
, and White is highly unlikely to break down the blockade; or, even better, 39...Qc8!
40.Bxd5
(40.Rxa6
Bxd4
41.Qxc8
Rxc8
42.Bxd5
Kf8=
) 40...b4
41.Qd2
Bb7
42.Nb3
Bxd5
43.Qxd5
Bb6
, and Black is fine.]
40.Qa5
Once again I barely made the move within the time limit - but White is winning again. In fact, even [40.Nc6
was probably enough: 40...b4
41.Nxe7+!
Qxe7
42.Qc6
Bb7
(42...Bc4
loses on the spot: 43.Ra7!
Qxe6
44.Qc7
) 43.Qb6!
, and the bishop is still causing Black grief: 43...Ra8
(43...Bc8
44.Ra7
Qxe6
45.Qc7+-
) 44.Rc1!
Rc8
(44...Ra6
45.Qc7
Qxc7
46.Rxc7
Bc6
47.e7
Kf7
48.Rxc6+-
) 45.Rxc8+
Bxc8
46.Bxd5
, and there is no good defense against 47.Qc6]
40...Bc5
Desperation. [40...Ra8
was more tenacious: in fact, White would have to get creative in order to avoid some opposite-colour endgames which might be very difficult to win. 41.Rc1!
(41.Nxb5
Rb8!
42.Nc7
Bb4
43.Qxd5
Qxc7
44.Rxa6
Be7
is a prime example: Perhaps this position is hard to hold in practice, but White would certainly have to sweat winning it.) 41...Rc8
(41...Bd8
42.Qa2
Rc8
43.Rxc8
Qxc8
44.Bxd5
Be7
45.Nc6
looks grim; computer-like 41...Kh8
does not hold either: 42.Nc6
Qc8
43.Qd2!
(but not 43.Bxd5
Bb7
44.Qxb5
Kg7
, and Black will hope once again) 43...Ba3
, and now White has to find an extremely beautiful idea: 44.e7!
Bxc1
45.Nd8!
Qd7
46.Qxd5
Qxe7
47.Qxa8
Kg7
48.Bd5
, and White wins the long-suffering bishop yet agian) 42.Rxc8+
Qxc8
43.Bxd5
Kf8
(43...Qc1+
44.Kh2
Qb2
45.Nc6
Qxf2+
46.Bg2
) 44.Nc6
, and White should win]
41.Nb3
Now White's task is easy.
41...Bxf2+
42.Kxf2
Qa7+
43.Kf1
Qe3
44.Qxa6
Qxb3
45.Qa3
[45.Qc6
was cleaner: 45...Qc4+
46.Qxc4
dxc4
47.Bc6
]
45...Qc4+
46.Kg1
b4?
[We both missed that after 46...Qd4+
47.Kh2
Qe5!
Black will be able to collect the pawn on e6. White is still winning, but it would take a while.]
47.Qa7
Now it's all over - there are no checks, and 47...Rxe6 allows mate.
47...Kh8
48.Qd7
Qc5+
49.Kh2
Qe7
50.Bxd5
Black resigned. 1-0