Tuesday, July 7, 2015

A Discussion Continued

After this game was finished, I checked with The Database, and discovered that my opponent and I had "discussed" this opening (over the board) several times previously.

In addition, before I post a game on this blog, I usually "discuss" it with either Houdini 3 or Stockfish 6, hoping for enlightenment.

This time, however, my two electronic "friends" did not always agree in the early stages, occasionally sounding like a couple of know-it-all kibitzers, leaving a current assessment of the opening lines a bit unsettled.

Ironically, the game was tipped by the endgame play of both me and my opponent - as HAL 9000 once said, "It can only be attributable to human error."

perrypawnpusher - michon

blitz, FICS, 2015

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nd4 4.Bxf7+ 

The Blackburne Shilling Jerome Gambit.


My opponent declined the piece a couple of times, a couple of years ago, with 4...Ke7, but without success: perrypawnpusher - michon, blitz, FICS, 2013 (1-0, 8) and perrypawnpusher - michon, blitz, FICS, 2013 (1-0, 41). 

5.Nxe5+ Ke8 

Instead, 5...Ke6 was seen in perrypawnpusher - michon, blitz, FICS, 2014 (1-0, 23); while 5...Ke7 was seen in perrypawnpusher - michon, blitz, FICS, 2013 (1-0,18). 


This move was preferred by Houdini 3 after some deep thought.

Some human input: In over 70% of the over-2,100 relevant Blackburne Shilling Jerome Gambit games in The Database, White chose 6.Qh5+, scoring 60%.

Instead, Stockfish 6 preferred 6.c3.

Again: In 21% of the relevant Blackburne Shilling Jerome Gambit games in The Database, 6.c3, was the people's choice, scoring 55%.

I checked, and noticed that 6.c3 was first mentioned on this blog in "Blackburne Shilling Jerome Gambit", 7 years ago.

Interestingly, an earlier version of Stockfish expressed a preference for 6.c3 about 5 years ago.

The choice of 6.Qa5+ or 6.c3 still seens to be an "Unfinished Discussion" (posted 3 years ago), although it seems that 6.c3 may have appeared in more blog posts than 6.Qh5+.

Here are a couple of examples of the 6.c3 Nc6 7.Nxc6 dxc6 8.d4 Qh4 line suggested by Stockfish 6: 9.0-0 (9.g3 Qxe4+ 10.Qe2 Qxe2+ 11.Kxe2 Bg4+ 12.f3 Bf5 13.Re1 Bxb1 14.Kf1+ Be7 15.Rxb1 Kd7 16.Re2 Re8 17.Be3 Nf6 18.c4 Rhf8 19.b4 Ng4 20.Kg2 Nxe3+ 21.Rxe3 Bxb4 22.Rxe8 Rxe8 23.Rxb4 Re2+ 24.Kh3 Rxa2 25.Rxb7 Ra4 26.Kg4 Rxc4 27.Rxa7 Rxd4+ 28.f4 c5 29.Ra8 c4 30.Rh8 c3 31.Rxh7 c2 32.Rxg7+ Kc6 33.Rg6+ Rd6 34.f5 c1Q 35.Rxd6+ cxd6 36.f6 Qc4+ 37.Kf5 Qf1+ 38.Ke6 Qh3+ 39.Ke7 Qd7+ 40.Kf8 d5 41.f7 d4 42.Kg8 Qg4+ 43.Kh8 Qh5+ 44.Kg8 Qg6+ 45.Kf8 d3 46.Ke7 Qxf7+ 47.Kxf7 d2 48.h4 d1Q 49.h5 Qxh5+ White resigned, marinrouge - homerg, FICS, 2002) 9...Ke7 10.Nd2 Nh6 11.Nf3 Qh5 12.Qd3 Kd8 13.Bg5+ Be7 14.Bxh6 gxh6 15.Ne5 Bd6 16.Qg3 Be6 17.f4 Bxe5 18.fxe5 Rg8 19.Qf3 Bg4 20.Qf7 Qxf7 21.Rxf7 Kc8 22.Kh1 b6 23.Raf1 Kb7 24.Rxh7 h5 25.Rff7 Rac8 26.d5 cxd5 27.exd5 Be2 28.d6 Rgf8 29.Rxc7+ Rxc7 30.Rxc7+ Ka6 31.h3 h4 32.Kh2 Bb5 33.Rg7 Bc6 34.e6 b5 35.d7 Rf6 36.Re7 Bd5 37.d8Q Rg6 38.Qc8+ Ka5 39.Qc7+ Ka4 40.Rg7 Rxg7 41.Qxg7 a5 42.Qd4+ b4 43.cxb4 Black lost on time, dzomba - VendettaA, lightning [!], FICS, 2006. 

6...g6 7.Nxg6 


Here we have another dispute.

Stockfish 6 prefers 7...Nxc2+, suggesting that after 8.Kd1 hxg6 (8...Nxa1? 9.Nxh8+ Ke7 10.Qe5#) 9.Qxg6+ Ke7 10.Qg5+ (Houdini prefers 10.Kxc2 with an edge for White) 10...Ke8 11.Qg6+ Ke7 12.Qg5+ White draws by repeating the position (Houdini prefers 12.Kxc2 with an edge for White).


A number of years ago I got away with the weaker 8.Qxh8?! after 8...Nh6 when 8...Nxc2+, instead, would have led to an edge (Houdini) or an advantage (Stockfish 6) for Black after 9.Kd1 Nxa1 10.Qxg8 in perrypawnpusher - tampajake, blitz, FICS 2009 (1-0, 12).

8...Ke7 9.Qg5+ Nf6

Here Stockfish 6 says Black has the advantage, while Houdini 3 says the position is equal.

Both prefer the text to 9...Ke8  of perrypawnpusher - adamzzzz, blitz, FICS, 2009 (1-0, 39). 


Better than 10.e5 of perrypawnpusher - JokeritT, blitz, FICS, 2010 (1-0, 28).


Stockfish 6 slightly prefers 10...Ke8 to the text, while Houdini 3 thinks 10...d6 is better. 

11.Qxd4 Be6 

Here, Houdini 3 likes White, while Stockfish 6 prefers Black.

Of course.

For the record, I have also seen:

11...c5 in perrypawnpusher - foreverblackman, blitz, FICS, 2010 (1-0, 21) and perrypawnpusher - vlas, blitz, FICS, 2010 (0-1, 23);

11...Qd7 in perrypawnpusher - theferno, blitz, FICS, 2010 (1-0, 28);

and 11...Bg7 in perrypawnpusher - Raankh, blitz, FICS, 2010 (1-0, 49). 


Also possible is 12.f4 c5 when again Houdini 3  likes White, while Stockfish 6 likes Black.

12...Bh6 13.d3 Bxc1 14.Rxc1

Both Stockfish 6 and Houdini 3 suggest 13...c5 first, to displace the White Queen.  

White has four pawns for his sacrificed piece. Even if the position is "objectively" even, his prospects are good at club play level.

The next few moves show that both players are blind to some tactical possibilities, however.

14...Qg8 15.g3 Bh3 16.Nd5+?! Nxd5 17.Qxd5? Qxd5 18.exd5 Rae8 

Both players overlook the fork 18...Bg2! when Black can grab a pawn and increase Kingside pressure after 19.Rg1 Rxh2 20.c4 Rf8.

19.Kd2 Kd7 20.c4 Re7 21.f4 Rhe8 22.Rce1 Rxe1 23.Rxe1 Rxe1 24.Kxe1

Both computer programs see the position as equal.

What follows goes along with suggestion that a bad plan is better than no plan at all - which is a nice way of saying that my bad idea worked, while my opponent missed some good ideas.

24...Bg4 25.Kf2 Ke7 26.Kg2?! Kf6?! 

Black is worried about stopping the Kingside pawns, when he probably should look at using his Bishop to capture a couple of center pawns: 26...Be2 27.d4 Bxc4 28.a3 Bxd5+ 29.Kf2 b6. After this, Black can look to create his own passed pawn, which should counter White's activity.

27.h3 Bd7?!

Again, 27...Be2 was to be preferred. Black's Bishop has to cut down the number of enemy pawns before they become a real problem. 

28.g4 Bxg4

This is tantamount to resignation. White can now develop play on both wings. At least one passer is bound to get through.


29...a6 30.Kf3 c5 31.dxc6 bxc6 32.Ke4 Kg6 33.d4 a5 34.b3 Kf6 35.a3 Kg6 36.b4 

Black resigned

No comments: