This is possible because 4.Nb5? is simply met by 4...Qa5+, forcing 5.Nc3, after which 5...cxd4 gives Black a large advantage.
Instead White normally plays 4.e3 (4.dxc5?! d4), supporting the d4 pawn and opening a diagonal for White's light-square bishop.
Semetey Tologontegin (IM 2369) - Oleg Korneev (GM 2511)
Cheliabinsk Panchenko Memorial 2019
4...cxd4 5.exd4 a6
5...e6?! is a natural-looking move that has ensnared some very strong players, for example Baadur Jobava (GM 2669) - Igor Kovalenko (GM 2668), Eurasian Cup Blitz (Almaty, Kazakhstan) 2016, continued 6.Nb5 Na6 7.c3!? Be7 8.Nf3 0-0 9.Bd3 Bd7 10.a4, when the Latvian GM lost patience with the annoying knight: 10...Bxb5?! (Stockfish10 and Komodo10's 10...Nh5 11.Be3 Nb8, planning ...a6, seems a better try) 11.axb5 Nc7 12.0-0 Bd6 13.Ne5. White's bishop-pair and queenside pressure gave him a large advantage (1-0, 61 moves).
However, the fact that Black plays ...a6 anyway is one reason why 3...a6 is popular at the highest levels (see parts two and three).
6.Bd3!?
Overwhelmingly more popular is 6.Nf3, but the text has been Jobava's choice and has also been played by Magnus Carlsen.
6...Nc6 7.Nge2
Carlsen played the eccentric-looking 7.Nce2 in a 2016 rapid game.
The text avoids a pin from Black's light-square bishop, which suddenly looks short of decent squares.
7...Bg4!?
Ruslan Ponomariov shut his light-square bishop in with 7...e6 against Jobava in a 2016 Olympiad loss.
The problem with the text is that the bishop becomes a target for white kingside expansion.
8.f3 Bh5 9.h4 Bg6 10.g4 e6 11.h5 Bxd3 12.Qxd3
Taking with the c pawn, as seen in Carlsen - Nakamura in part three of this series, is not appropriate here as White would not have a mobile centre, and the isolated d pawns would be weaknesses.
12...h6 13.0-0-0
Better than the Kf2?! seen in a similar position in part two.
One of the attractions of the Jobava-Prié for an attacking player is that White often gets to castle long in relative safety, while Black is unsure whether to castle short, where he faces an imminent pawn-storm, or to keep his king in the centre, when his pieces may lack coordination.
13...Bd6 14.Qe3 Na5 15.b3
 |
| How should Black proceed? |
*****
*****
*****
*****
15...Rc8!?
Korneev is the first to vary from Jobava (GM 2646) - Liviu-Dieter Nisipeanu (GM 2657), Spilimbergo (Italy) 2018, which saw 15...Bxf4 16.Nxf4 Rc8 17.Nce2 Qd6 18.Nd3 0-0 19.Kb1, when the engines prefer White but in the game Black's queenside attack triumphed.
16.Kb1 Nc6 17.Bxd6 Qxd6 18.Nf4 b5 19.Nd3 Nd7 20.Ne2 Ne7 21.c3 a5 22.Rc1 a4!?
Komodo10's choice, but it allows White to lock the queenside.
23.b4 Nb6 24.Nf4 Nc4 25.Qe2 Na3+ 26.Kb2 Nc4+ 27.Ka1
The IM declines to see if his GM opponent is happy with a draw - Tologontegin believes he is better, or at least that he has the safer king.
27...Nc6 28.Rhg1 Qe7 29.Rce1 Rd8 30.Qf2 0-0
Black at last castles, and so gets his king's rook into play. But unless he can engineer a sacrificial breakthrough on the queenside, which seems unlikely, the best he can apparently hope for, barring a blunder from White, is a draw.
31.h3 e5!?
Seeking central counterplay before White organises a kingside attack.
The engines suggest 31...Qf6, but after 32.f4 the storm clouds are gathering.
32.g5 hxg5 33.Rxg5 e4?
Komodo10 reckons 33...Qe6 holds, but Stockfish10 continues 34.dxe5!, when the engines agree White is much better, eg 34...Qxh3 35.Nf4 Qc8 36.Qg3 Kh8 37.Ng6+! Kg8 (forced), and now the simple 38.Nxf8 is good enough for a large advantage, although Stockfish10 reckons 38.Nh4!? is even stronger.
Perhaps best is Stockfish10's suggestion of 33...Kh8, although White is clearly better after, for example, 34.Qh4.
34.fxe4 dxe4 35.Qg2?
Very strong is the engines' 35.h6 g6 36.Rxe4! as Black gets mated after 36...Qxe4? 37.Qf6 Qh1+ 38.Rg1 etc. Also winning is 35.Reg1.
35...exd3?
The engines' 35...Qf6 seems to hold as 36.Rf1? runs into 36...Qh6, when Black threatens a fork on e3 (as well as capturing the d3 knight), and 37.Qxe4 is met by another fork, 37...Nd2.
36.Rxe7 Nxe7 37.Rxg7+ Kh8 38.Qg5 Nf5?
This lets White mate, but the engines' 38...Rd6 offers only limited hope.
39.Qxf5! Kxg7 40.h6+ Kxh6 41.Qf6+ Kh7 42.Ng5+ Kg8 43.Qh6 1-0
Quite a demolition job by the IM, and a fine advert for the Jobava-Prié, with the only obvious blemish being the error at move 35.
No comments:
Post a Comment