Friday, 3 October 2014

King's Gambit 3.Nf3 g5 lines revisited- 4.h4 and 4.Bc4

After 1.e4 e5 2.f4 exf4 3.Nf3, Black's most theoretically critical response is to hold onto the f4-pawn by playing an early ...g7-g5.  The most popular way of doing so is with the immediate 3...g5 although there is also a strong case for the Fischer Defence (3...d6 intending 4...g5) which I intend to take a closer look at in the near future.  I may also need to re-examine 3...Nc6 intending 4...g5, which I used to use quite a lot in my own games before being put off by certain lines where White plays a quick d2-d4-d5.  However, 3...Nc6 remains my favourite response to 3.Bc4.

For now my revisits to these lines have focused upon 4.h4 and 4.Bc4.

Traditionally most often recommended for White during the 1990s was the Kieseritzky Gambit, 4.h4 g4 5.Ne5.

John Shaw has a very extensive analysis of some attractive possibilities for White in the line 5...Nf6 6.Bc4 (best, instead 6.d4 d6 7.Nd3 Nxe4 is difficult for White) 6...d5 7.exd5 Bd6, stemming from some of David Flude's ideas.  I have chosen to just give a few sample lines- those interested in a very extensive analysis would be advised to get his book.  

However, it hasn't added to my enthusiasm for the line because I have seen nothing to change my assessment that 5...Qe7, envisaging a quick queen trade on the e-file, leads to a rather sterile equality, where if anything Black's chances of getting a slight "nibble" (as Mark Hebden used to call it) are greater than White's.  I have seen further evidence of why 5...d6 6.Nxg4 Nf6 7.Nxf6+ Qxf6 is regarded as drawish at high levels of play, though I think that at club level, there is plenty of scope for both sides to go wrong early in the game, so I don't think 5...d6 should bother club players too much.  Finally, there is 5...Nc6!? intending 6.d4 Qf6, which is a good way for Black to get unbalanced and equal play and a fair share of the attacking chances.

Meanwhile the Allgaier Gambit, 5.Ng5, is still looking somewhat dubious, though I can see scope for some players to have a lot of fun with it in casual games, since White does get a strong attack if Black plays inaccurately.  Both 5...f3!? and 5...h6 are looking strong for Black.  Fans of this approach are well-advised to try the Quaade Gambit (4.Nc3) with the aim of reaching the Hamppe-Allgaier Gambit following 4...Nc6 5.h4 g4 6.Ng5, where the inclusion of Nc3 and ...Nc6 improves White's chances.

Then there is 4.Bc4, which often leads to a Hanstein Gambit.

The piece sacrifice line 4...g4 5.0-0 gxf3 6.Qxf3 is still looking quite attractive for White, as long as White meets 6...Qf6 with 7.d3, rather than the further pawn sacrifice 7.e5, which is looking dubious.

However, Black does better to set up the solid f4-g5-h6 pawn chain.  John Shaw has suggested a couple of improvements over my previous analysis of the line 4.Bc4 Bg7 5.d4 d6 6.0-0 Nc6 7.c3 h6 8.Qa4 (which has been recommended and analysed extensively by Stefan Bücker).  The idea of that line is to encourage 8...Bd7 so that 9.Qb3 cannot be met by 9...Qd7.  The main idea is that after 8...Bd7 9.Qb3 Na5 10.Bxf7+ Kf8 11.Qa3 Kxf7 12.Qxa5, Black plays 12...Ne7, completing development, rather than hitting out at White's centre with 12...c5.  While I think Shaw is too dismissive of the line from White's perspective (White does have a strong pawn centre and a safer king to offset Black's better development and bishop-pair) a closer look at the line does suggest that if anyone stands slightly better, it is probably Black.  Alternatively, White can hack around with h2-h4 and b2-b4-b5, raising questions about where Black should put the king, since it is prone to attack on either side of the board, but again with accurate play Black should be able to get the better chances.

A reasonable alternative for Black is 4...Nc6, which often arises from the Bishop's Gambit: 3.Bc4 Nc6 4.Nf3 g5, but it will usually transpose into the same variations.

Thus I am no longer convinced that 4.Bc4 leads to dynamic equality with best play- Black may well be able to eke out a theoretical edge.  However, as I know from experience of playing the black side of these lines, it is far from easy for Black to demonstrate this at club level.

The next area to be revisited will be the Quaade and Rosentreter Gambits with 4.Nc3 and 4.d4, respectively, which I tend to favour from the white side.


  1. Brabo has written on the Quaade recently:

  2. I "hijacked" your most recent blog, about the Alekhine-Chatard Attack, and made a couple of comments regarding the Petroff Defence and potential transpositions to the Scotch Gambit. Since then, I have begun to read up on your blog, bookmarking posts that I will go back and read more carefully later. At this point, I would like to make a couple of comments.

    First, you often write about how you do not have time to update the blog and your associated opening site as often as you would like to, but in my opinion this is far from a problem. In my view, maintaining your high quality is far more important than updating frequently, and I am sure that most readers agree with me on that.

    Second, regarding the topic of this blog, I would like to direct your attention to a high-level King's Gambit game played this summer. (Since I have not yet read everything on your blog, I am not sure that you have not seen it, but I thought that I should at least recommend it.)

    In this year's Swedish Championship, GM Nils Grandelius had a one-point lead before the last round, when he was facing GM Emanuel Berg as black. Berg had to defeat Grandelius to share the first place and qualify for the tiebreak, and what did he play when he had to win the game? The King's Gambit!

    The game is available on and apparently one mistake by black was enough to give white an advantage that lasted into the endgame. I thought this might be interesting as both players are rated well above 2500 (and Grandelius has even surpassed 2600).

    Kind regards,

    1. I hadn't considered 11...Nd7, but it had not been played before on the database, so it may be a novelty. It turns out that John Shaw covers 11...Nd7 in his book and the game exactly followed his recommended line, up to 18.c4, which he gave as clearly better for White. I note that if White plays 18.Bh6 (18.Bd2 is similar) then White ends up with bishop and knight for rook and pawn following 18...Ne5 19.Rf5 Rxd5 20.Rxh5 Nf3+ 21.gxf3 Rxh5. That variation is better for White too, but I'm not convinced that White's advantage is larger than after 18.c4. Regardless, this looks like a distinctly inferior line for Black. 11...0-0 is better, but I think Black's best variation in the 5...Nf6 variation of the Kieseritzky is to deviate with 9...0-0. White can still look forward to an equal and complicated/unbalanced game then.