Back To The Drawing-Board (part three)?

IT is interesting to see what has been happening between strong players in the line 1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nxf7!?, which is generally known as the Cochrane Gambit after 19th-century Scottish barrister John Cochrane.
Here is a relatively recent high-level game at a normal time-control.
Aleksei Pridorozhni (2589) - Semen Khanin (2485)
Voronezh (Russia) 2018

4...Kxf7 5.d4

Lasha Janjgava in The Petroff (Gambit 2001) says 5.Bc4+ is "suspect" in view of 5...d5! (less popular than 5...Be6 and 5...Ke8 in ChessBase's 2020 Mega database, but apparently a lot stronger) 6.exd5 Bd6 when White has three pawns for a knight, but the black king will find safety on g8 without blocking the h8 rook.

Janjgava says of the 4.Nxf7!? line in general: "Theoretically speaking, there isn't much to commend the sacrifice, but White's central control is quite good and Black has considerable long-term difficulties in finding a safe haven for his king."

5...c5!?

This somewhat odd-looking countergambit was unknown when David Hooper wrote A Complete Defence To 1 P-K4 - A Study Of Petroff's Defence (Pergamon 1967). It seems to have been first played at a high level in the early 1980s, the idea being to diminish the power of the white centre.

6.Nc3!?

After the obvious 6.dxc5 the analysis engines Stockfish11 and Komodo11.01 reckon Black gets a big advantage with 6...d5!? 7.e5 Qe8.

6...Nc6

Both players have lots of choice early on in this variation as there are plenty of byways where theory has not made a definite pronouncement.

7.Bc4+

Closing lines with 7.d5?! Nd4 is probably wrong. White needs a fluid centre to create threats before Black can get his king to safety and win comfortably on material.

7...d5!?

Offering a third pawn, at least temporarily, but the alternatives were not attractive as 7...Be6?? drops a piece while 7...Ke8 does nothing for Black's development or king safety.

8.Nxd5

An improvement on 8.exd5? Nxd4, when Black had the sensitive e6 square well-covered in Bolivar Ribeiro Gonzalez (2317) - Simão Poscidônio Dias (2029), Rio de Janeiro Zonal 2011 (0-1, 50 moves).

8...Be6

One month later the game Miroslav Marković (2470) - Miodrag R Savic (2545), Novi Sad 2018, saw 8...Na5 9.Nxf6+ Nxc4 10.Nd5 cxd4 11.Qxd4 Nb6 12.0-0 Nxd5 13.c4!? Qf6 14.e5 Nf4 15.Bxf4 when White's three pawns for a piece proved insufficient compensation (0-1, 39 moves).
9.Bg5 cxd4 10.0-0 Be7 11.Nxe7 Qxe7 12.Bd3

White has two pawns and the bishop-pair for a piece and Black's king is still not completely safe

12...h6 13.Bh4 g5 14.Bg3 h5 15.f4 h4 16.Be1
Both players have been going for it, but here the engines prefer 16.fxg5!? when 16...hxg3 17.Rxf6+ looks dangerous for Black.
16...Bg4
Komodo11.01 suggests developing the queen's rook with 16...Rag8, while Stockfish11 likes 16...Ke8!?, but the consequences of both moves are unclear.
17.Qd2 gxf4 18.h3 Rag8!? 19.hxg4 Rxg4 20.Rxf4 Qd6 21.Rf5 Ne5
At the cost of giving back the piece and emerging a pawn down, Black has taken over the initiative.
22.Qb4 Qe6??
A blunder. Black had to exchange queens, when the position remains unclear.
23.Qb3?
The simple 23.Qxb7+ is strong but the engines reckon White's advantage is overwhelming after the quiet developing move 23.Bd2, when White is ready to play Raf1 as well as threatening the loose pawns on b7 and d4.
23...Qxb3 24.axb3 Ke6 25.Ra5!
Not 25.Rxa7? Nxe4.
25...Nc6?
Black's disadvantage is much smaller after 25...Nfd7, according to the engines.
26.Bc4+ Ke7 27.e5 Nd7 28.Rb5
Also very strong is 28.Rf7+.
28...h3 29.g3 h2+ 30.Kh1 Re4 31.Bd2 Ncxe5
There is nothing better, according to the engines. As Janjgava warned can happen, Black has failed to find a long-term safe-haven for his king.
32.Bb4+ Ke8 33.Rxb7 a5 
Not much better is 33...Nxc4 34.bxc4.
34.Bxa5 Nxc4 35.bxc4 Re2 36.Ra7 Rh3 37.Bb4 Rxc2
Or 37...Kd8 38.Rg5.
38.Ra8+ (1-0, 54 moves)
I imagine much of the game was played in time trouble, which gives added weight to Jangjava's observation that "a tactically minded player might be tempted by the sacrifice" 4.Nxf7!?

New Opening That's All The Rage

PLAYED on board five last night for Battersea 2 against British Bangla in division two of the London League.
Nasir Rizvi (182) - Spanton (170)
'Barry Attack'
1.d4 d5 2.Nc3 e6 3.Bf4
3.e4 transposes to the French Defence. The text - first played in 1856 in Calcutta - introduces a cross between the Veresov (1.d4, 2.Nc3 and 3.Bg5) and the New London (1.d4 and 2.Bf4). One of its names is the New Veresov. I call it the Barry Attack, but this name is usually restricted for 1.d4, 2.Nc3 and 3.Bf4 against a black set-up involving an early kingside fianchetto. Another name is the Jobava-Prié Attack.
After the game, five Battersea players went to a nearby Wetherspoon. It turned out all of us, or perhaps it was just four of us, had faced 1.d4, 2.Nc3 and 3.Bf4 either that night or the night before. Clearly something is going on.
3...Bd6
The main move is 3...Nf6, as played in Moheschunder Bannerjee - John Cochrane, Calcutta 1856 (0-1, 37 moves), but the text has a better percentage score in ChessBase's 2020 Mega database.
4.e3!?
White's score leaps to an impressive 61% after this, and it can easily be seen that the text saves a tempo over 4.Bg3 and 4.Bxd6 Qxd6.
4...Bxf4
Trying to punish White's play. Black cannot necessarily delay this decision as 4...Nf6 has been met with 5.Qf3!? by Romanian international master George-Catalin Ardelean.
5.exf4 Nf6
Logical-looking is 5...Qd6!? but it is not liked, at least at first, by Stockfish10 and Komodo10, who continue 6.Qg4 Kf8 7.Qg3 with an unclear position that Stockfish10 reckons may slightly favour Black, but Komodo10 reckons favours White. As the saying has it, tests are needed.
6.Nf3 Qe7?!
Now 6...Qd6 can be met by 7.Ne5, which occurred in Richárd Rapport (2698) - Miklos Nemeth (2451), Zalakaros (Hungary) 2017 (1/2–1/2, 36 moves).
The problem with the text is that the e pawn will be pinned against the queen when a white rook lands on e1.
7.Bd3 Bd7?!
7...c5 8.dxc5 Qxc5 9.0-0 Nc6 10.Nb5 seemed slightly better for White in Igor Miladinović (2555) - Robert Markus (2625), Serbian Championship 2012, but was agreed drawn just eight moves later.
The text is what I suspect Garry Kasparov would call a half-move, ie one that is vaguely useful but does not get to the heart of the position. I was reluctant to play 7...0-0 as I feared White would castle long and launch a kingside pawn-storm.
8.0-0 0-0 9.Re1 c5 10.dxc5
10.f5 cxd4 11.fxe6 dxc3 12.exd7 Qxd7 looks fine for Black.
10...Qxc5 11.Ne5 Nc6
11...Qb4?! looks very greedy, but is maybe playable.
12.Re3!?
The engines prefer a slower build-up with, for example 12.Qd2.
12.g4!? is sharp. The engines continue 12...Qb4!? 13.Nxd7 Nxd7 14.Qf3 with a menacing-looking position for White, but the engines like Black.
12...d4!?
Forcing, but the engines prefer 12...Qb4 13.Rh3 g6 14.Qc1!? with an unclear position.
13.Ne4 Nxe4 14.Nxd7 Qd5
Not 14...Nxf2? 15.Nxc5 Nxd1 16.Rh3.
15.Bxe4 Qxd7

How should White proceed?

*****
*****
*****
*****
*****
16.Bxh7+?
Unsound.
White has a slight edge after 16.Rd3, according to the engines.
16...Kxh7 17.Rh3+ Kg8 18.Qh5 f6 19.g4 Qe8 20.Qh7+ Kf7 21.g5 fxg5 22.fxg5 Ke7!?
A good practical choice, although the engines narrowly prefer 22...e5.
23.Re1
Objectively best is 23.Qxg7+!?, but 23...Qf7 forces queens off as Black threatens f2.
23...Qf7 24.f4 Rh8 25.Qd3 Rxh3 26.Qxh3 Rf8 27.Qa3+ Kd7 28.Qb3 Kc7!?
Another practical choice.
The game finished:
29.Qxe6 Qxf4 30.g6 Rf6 31.Qe8 Qg4+ 32.Kh1 Rf2 33.Re7+ Kb6 34.Rxb7+ Kxb7 0-1
My updated Battersea statistics for 2019-20
Event*..Colour...Grade...Opponent's Grade...Result
CLL..…...B....…..168...………169...……….....W
CLL...…..B...…...168...………196...…………..D
CLL...…..W...…..168...………176...…………..L
LL....……W...…..168...…....…175...………….D
CLL...…..B...…...168...………192...…………..D
CLL...…..W...…..168...………181...…………..L
LL...…….W...…..168...………168...…………..L
EC...…….B...…..168...………175...……….….L
CLL...…...B.........170...………172...…………..L
LL...…….W...…..170...………183...…………..D
LL............W........170...…........180......……......L
LL....…….B...…..170....……...182...………….W
Overall this season for Battersea I have scored +2=4-6 for a grading performance of 162.
In season 2018-19 I scored +12=12-13 for a grading performance of 169.
In season 2017-18 I scored +10=8-9 for a grading performance of 175.
*CLL: Central London League; LL: London League; EC: Eastman Cup.

Morphy v The Evans (part nine)

JULES Arnous de Rivière had the chance to play a third Evans' Gambit against Paul Morphy in Paris in 1858, this time in consultation with another French master, Paul Journoud.
Journoud & Rivière - Morphy
1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.b4 Bxb4 5.c3 Bc5 6.0-0 d6 7.d4 exd4 8.cxd4 Bb6
Once again reaching the Normal Position.
9.Bb2 Nf6 10.Nbd2 0-0
So far the game has followed Rivière's last Evans' Gambit against Morphy, but here the partners play a major improvement over that game's 11.h3?!
11.e5 dxe5
Still the main reply today but, as stated in part eight of this series, the engines Stockfish10 and Komodo10 suggest the seeming-novelty 11...Nh5!?
12.dxe5 Nd5 13.Ne4 Be6 14.Nfg5 h6!?
This appears to be the best move in the position, even though it weakens the light squares around Black's king.
14...Nf4 was tried six years previously by Moheschunder Bannerjee ("Mohishunder Bonnerjee" in ChessBase's 2019 Mega database) in two games against John Cochrane in Calcutta. Cochrane replied 15.Bxe6?! both times and won the games, but Black seems to stand well after 15...Nxe6 or 15...fxe6. The engines reckon much better for White is 15.Qc2!, continuing 15...Bf5 16.Nxf7! Rxf7 17.Bxf7+, all of which was played in Mikhail Chernov (2188) - Andrei Kuznesov (2288), Siberia Championship 2012. That game continued 17...Kxf7 18.Nd6+! Kg8 19.Nxf5, when White was the exchange up and had pressure against Black's king. Worse in this line would be 18...cxd6? 19.Qxf5+ Kg8 20.Qxf4, when Black is again the exchange down but with less active play. It seems Chernov (b1994), now an FM, had prepared well.
15.Nxe6 fxe6 16.Qg4

Position after 16.Qg4 - as GM Nigel Davies might put it, Black's king has the wind blowing in his hair

*****
*****
*****
*****
*****
16...Kh8!
The 'obvious' 16...Qe7 fails to 17.Nf6+! Kh8 18.Bxd5 exd5 19.Qg6! gxf6 20.exf6 Qh7 (20...Qf7? 21.Qxh6+ is a pawn-down version of the same line) 21.Qg7+ (21.f7+?? d4) Qxg7 22.fxg7+ Kg8 23.gxf8+ Rxf8, all of which was seen in Cochrane - Bannerjee, Calcutta 1852. White is clearly better, being the exchange up for a pawn, but the game has a silly finish in Mega19 (and chessgames.com) in which White resigns despite maintaining his material advantage.
17.Rad1!
An improvement on yet another Cochrane - Bannerjee, Calcutta 1852, game, which saw 17.Qxe6 Nf4 18.Qg4, when 18...Bd4! wins for Black, eg 19.Rfd1 (19.Bxd4 loses a piece to 19...Qxd4) Nxe5 20.Rxd4 Nxg4 21.Rxd8 Raxd8, with Black up the exchange and a pawn.
17...Rf4?
Morphy loses his way in the complications. 17...Qe7 was better. White can win back his pawn with 18.Bxd5 exd5 19.Rxd5, but Black is slightly better after 19...Rad8 thanks to White's weak e pawn.
18.Qxe6 Rxe4 19.Bxd5 Nd4 20.Rxd4 Rxd4 21.Bxd4 Bxd4 22.Re1?!
White can safely grab a pawn with 22.Bxb7 Rb8 23.Be4.
22...Qg5 23.Bf3
It seems Bxb7 was still possible, but Black gets a strong initiative with 23...Rf8.
23...Rf8 24.Qc4 c5?
Morphy's instinct was always against pawn-grabbing, but the text lets White's e pawn advance with menaces. Equal was 24...Bxe5 25.Bxb7.
25.e6 Qe7 26.Qa4
The engines reckon that stronger was centralisation with 26.Qd3, eg 26...Rf6 27.Qe4 b5 28.Qd5 b4 29.Qd7 Rf7 30.Qc8+ Rf8 31.Qxf8+! Qxf8 32.e7, winning. This line is by no means forced, but Black faces a difficult defence in any event.
26...Qh4 27.Qc2 Qe7 28.Qg6!
This is even better than transposing to the previous note with 28.Qd3, as after 28.Qd3 Black is not forced to play 28...Rf6.
28...Rf6
This is now forced, so White can transpose to the strong line given at move 26.
29.Qe4 Rf8
The engines come to prefer this to 29...b5.
30.g3
A horrible blunder would be 30.Qxb7?? Qxb7 31.Bxb7 Bxf2+ etc.
30...b5 31.Re2 b4 32.Bh5 a5 33.Bf7 a4 34.Qc2 Ra8 35.Qc4 Rb8 36.Qd5?
A mistake that allows serious counterchances. Better was 36.Bg6!, the point being that if 36...b3, as in the game, White has 37.Qxa4 b2 38.Bb1, leaving him a pawn to the good over the game continuation.
36...b3 37.axb3 axb3 38.Bg6 b2 39.Bb1 Rd8??
The position was dead-equal, but this uncharacteristic blunder by Morphy costs him the game.
40.Qf5
40.Qe4 also wins. Morphy must have seen 39.Bb1 as simply a defensive move, missing that it also sets up a possible battery against h7. Black would have been fine after 39...g6 or 39...g5.
40...g5
Best, but no longer good enough as 39...Rd8?? has given White a decisive tempo.
41.Qg6 Bf6 42.Qxh6+ Kg8 43.Qg6+ Bg7?!
Black's king is too weak for any defence to succeed, but perhaps a better practical try was 43...Kh8, eg 44.Kg2 c4 45.Re3 Rd3!? (45...Qg7 46.Qe4) 46.Bxd3 cxd3 47.Re1 d2 48.Rd1 Qb7+ 49.f3 Be7, but 50.h4 gxh4 51.gxh4 looks convincing enough, despite Black having two pawns on the seventh.
Another try is 43...Qg7, but after 44.e7 the best Black can do is 44...Rd1+ 45.Kg2 Qxg6 46.Bxg6 Bxe7 47.Rxe7 c4 (47...b1=Q 48.Bxb1 Rxb1 49.Re5 is a fairly-simple winning rook-and-pawn ending) 48.Kf3 c3 (the best try as 48...b1=Q 49.Bxb1 Rxb1 50.Re5 is another easy win) 49.Rc7 Rc1 50.Rb7, when Black's connected passed pawns have been neutralised, and White should win with reasonable care.
44.Qh7+
There was nothing wrong with the simple 44.Rxb2.
44...Kf8 45.h4!?
White is probably still winning after this, but it is less clear-cut than a patient line such as 45.Qf5+ Kg8 46.Kg2.
45...Rd1+
45...gxh4 46.Qf5+ Bf6 47.Rxb2 (47.gxh4? Rd1+ 48.Kg2 Rxb1! 49.Qxb1 Qg7+ gives Black serious drawing chances thanks to the exposed white king) hxg3 48.fxg3 Qg7 49.Qf3, and White remains a pawn up, albeit with mutually endangered kings.
46.Kg2 Bf6
46...gxh4 can be met by 47.Re4! Rd4 48.Re3.
47.hxg5 Bd4?!
The engines much prefer 47...Qxh7 48.Bxh7 Bxg5 49.Rxb2 Be7, but White has three connected passed pawns, and the presence of opposite-coloured bishops does not seem to be enough to save Black.
48.Qf5+ Kg8
Or 48...Ke8 49.Qf3 Rxb1 50.Qa8+ Qd8 51.Qc6+ Ke7 52.Qb7+ Kd6 53.e7 Qe8 54.Qb6+ Kd7 55.Qb5+ Kc7 56.Qxe8 Rg1+ 57.Kxg1 b1=Q+ 58.Kg2 Qb7+ 59.f3.
49.Qg6+ Kh8 50.Re4 Qb7 51.e7 1-0

Morphy v The ...e6 Sicilian

THE vast majority of Paul Morphy's opponents met his invariable 1.e4 with 1...e5.
But the Sicilian Defence was popular for much of the 19th century, and Morphy met one of its champions, the German master Louis Paulsen, in the first American Chess Congress.
This was held in New York in 1857 as a knockout tournament, and had the happy coincidence that the two best players met in the eight-game final.
In all four games in which Morphy had the white pieces, Paulsen replied with what is sometimes still called the Paulsen Variation of the Sicilian, and it is these four games I will be looking at in this series.
Morphy - Paulsen
Game 1
1.e4 c5 2.d4
Morphy varied between 2.d4 cxd4 3.Nf3 and the modern move-order of 2.Nf3 and 3.d4. He never played the Morra.
2...cxd4 3.Nf3 e6
3...e5!? is usually treated scornfully in opening publications, but it was successfully used by Howard Staunton in 1842 against John Cochrane, and is regarded by the analysis engines Stockfish10 and Komodo10 as giving an equal game.
4.Nxd4 Bc5!?
This has largely been superseded by 4...a6, 4...Nc6, 4...Nf6 and 4...Qb6, but is a favourite of the Russian GM, and former Karpov second, Vladimir Epishin.
5.Nb3
Morphy is credited by an anonymous annotator in ChessBase's 2019 Mega database as the inventor of this move, at least when used with Bf4 to highlight Black's weakness at d6. While this is not true - the plan was used by Tassilo von Heydebrand und der Lasa in a game seven years earlier - it may be fair to say Morphy's use of it in his match against Paulsen was a major factor in popularising the idea.
5...Bb6 6.Nc3 Ne7
Overwhelmingly the most-popular move today, but Heydebrand's opponent, August Ehrmann, preferred the modern-looking 6...a6.
7.Bf4 0-0
7...d5!? has been successfully used by the Filipino IM Paulo Bersamina.
8.Bd6 f5
Black's position looks rather restricted, but Paulsen was apparently relying on this pawn-break to create kingside counterplay.

How should White proceed?

*****
*****
*****
*****
*****
9.e5
Keeping the f file closed might seem a no-brainer, but Russian-born GM Sergei Tiviakov preferred 9.Qd2!? in a 1999 win over a 2294. I guess his thinking was that ...fxe4 is double-edged as White replies Nxe4 with hopes of his own quick kingside attack.
9...a6
Jim Plaskett - then an IM, now a GM - preferred 9...f4 in a 1983 draw against GM Sergey Kudrin.
10.Be2 Nbc6 11.0-0 Rf7
The game has transposed into the 1850 game Heydebrand - Ehrmann, which continued 12.a4 Ng6 13.Bh5, when the engines reckon Black could have freed his game with 13...Ncxe5!? 14.Bxe5 Nxe5 15.Bxf7+ Nxf7.
Morphy prefers to go for a kingside attack before Black can mobilise his queenside.
12.Kh1 f4!?
As the ChessBase annotator notes, this prevents 13.f4 and frees f5 for the black king's knight, but also frees e4 for White's pieces.
13.Ne4 Nf5?!
The engines prefer a ChessBase suggestion of targeting the e5 pawn by 13...Ng6, although they believe White has more than enough compensation for the fall of the pawn.
14.Bh5 g6?!
This weakens f6, which is why ChessBase and the engines prefer 15...Nxd6 followed by ...Rf8.
15.Bg4
Even stronger is 15.Bf3!, the point being that 15...Nxd6 can be met by 16.Nxd6 Rf8 (ChessBase suggests sacing the exchange with 16...Nxe5!?) 17.Bxc6, when White has a monster knight entrenched on d6.

Black is under pressure, but can you come up with a better move than Paulsen managed?

*****
*****
*****
*****
*****
15...Ng7?
As the ChessBase annotator laconically puts it: "Not easy to understand." Better was 15...Nxd6 16.Nxd6 Rf8, when the knight on c6 gives Black distracting pressure against e5.
16.Qf3 h5 17.Bh3 Qh4
Paulsen seems to have overestimated his attacking chances. However, the engines' suggested 17...Ne8, protecting f6, does not inspire confidence - Black's queenside pieces are largely locked out of the game.
18.Nf6+ Kh8 19.Qe4 Qg5 20.g3 f3
This pawn will be lost, but it is understandable that Paulsen did not want to play 20...fxg3 and open the f file for the white rooks.
21.Nd2! Bd8
Not 21...Qxd2?? 22.Qxg6.
22.Nxf3 Qh6 23.Rg1 Bxf6 24.exf6 Ne8
Not much better is 24...Rxf6. The engines give 25.Bf4 Qh7 followed by, for example, 26.Qe2 Nf5 27.Rad1 d5 28.Rge1 Bd7 29.Ne5, when it is hard to see Black holding out for long.
25.Bf4 Nxf6
Moving the queen is equally hopeless.
26.Qxc6 Qxf4 27.Qxc8+ (1-0, 40 moves)