King's gambit with Nf3

While the "refutation"of the King's gambit with Bc4 only arouse my curiosity, it was mainly the anti-dote against my favorite system of the King's gambit with Nf3 which worried me. I already play approximately 20 years the Fischer-defense. I am not aware of any anti-dote, at contrary as in recent years I more and more got convinced of the viability of blacks system. Besides we see also today top-grandmasters still choosing regularly for this setup.

The available excerpt on the official site of quality chess doesn't give a hint about the anti-dote but thanks to some reviews on chessvibes and Marsh Towers I was able to find out more. In the reviews is explained how white tries to get a favorable transposition to the Quaade-gambit via 5.g3.

1.e4 e5 2.f4 exf4 3.Pf3 d6 4.d4 g5 5.g3 !?

I have already played with blacks position after move 4, 5 official games and about 320 online blitz/bullet games but never I met 5.g3. On top a quick check with the engines let us understand that the idea is not easy to answer. Did John Shaw discover something or was this already known? I recently was able to look into the book and it stroke me that nowhere is mentioned from where 5.g3 originates. Not one example from practice is covered in the book. So the author invented the idea. No as I did find 5 games in the megadatabase  of which 2 white-games of the Dutch grandmaster Harmen Jonkman and 1 of the Russian grandmaster Vadim Zvjaginsev. Besides the latter player, Vadim is one of my favorite players whom I invariably replay his games when I accidentally come across them. Maybe the reader still remembers his introduction of 2.Na3 against the Sicilian in the Russian Superfinal of 2005. Something he still dared to repeat at least 8 times even against + 2600 players.

To ignore the history of the King's gambit is too bad (see my article manuals) but to present ideas as new in a book is at least dubious. You could claim that the earlier mentioned mega-database games were all lost by white so are irrelevant but a.f.a.i.k. an idea should be evaluated on its value and not only on the score (or the rating of the players, see blogarticle theory). On the other hand I have to admit that the author is the first to really initiate an attempt to popularize the idea. Loyal readers know from the blogartikel SOS that I am rather sensitive for that kind of details.

Enough about the origin of the idea as now we want surely to know how dangerous and interesting the idea is for practice. Beginning of June I started with a study of the idea lasting several weeks without looking in advance at what John Shaw tells. I was up to date with my analysis of the own played games so some time remained till Open Gent to make some extra opening-analysis as I did earlier e.g. for the Aljechin see article. By the way in an open tournament the chance is considerably bigger to meet opponents choosing for an experiment with e.g. the King's gambit and on top are also informed about the latest developments.

Not peeking to what the book says, has the advantage that one can look open-minded to the position which permits easier to find new ideas. This way I indeed found an interesting concept with f5 which checking later wasn't covered in the book.

[Event "KG Fischersysteem 5.g3 f5"] [Date "2014"] [White "?"] [Black "?"] [Result "*"] [ECO "C34"] [PlyCount "21"] 1. e4 e5 2. f4 exf4 3. Nf3 d6 4. d4 g5 5. g3 f5 $146 {(A cute idea not mentioned by JS but objectively it is not really better than the known alternatives.)} 6. Nc3 (6. exf5 Bxf5 7. Bg2 Nd7 8. O-O Qf6 9. Nc3 O-O-O 10. Nd5 Qg7 11. gxf4 $11 ) 6... Bg7 7. exf5 Bxf5 8. Bd3 Bxd3 9. Qxd3 Nc6 10. gxf4 gxf4 11. Bxf4 $11 *

I surely don't claim any advantage for black after f5 but it does seem very well playable. Somebody blindly following Johns book, can be forced quickly to think independently which in this explosive position isn't so funny.

Next I had a look to g4 which the engines recommend. We enter hereby the Quaada-territory. I spent a lot of time on the different sort of positions as this is very virgin territory and a lot of subtleties decide about what is and what is not playable.

[Event "KG Fischersysteem 5.g3 g4"] [Date "2014"] [White "?"] [Black "?"] [Result "*"] [ECO "C34"] [PlyCount "26"] 1. e4 e5 2. f4 exf4 3. Nf3 d6 4. d4 g5 5. g3 g4 {(Hereafter white gets the typical compensation that he expects.)} 6. Nh4 f3 {(JS does not talk about any deviations from the Quaade gambit.)} (6... Bh6 7. Bb5 c6 8. Bc4 Bg5 9. O-O Bxh4 10. gxh4 f3 11. Nc3 Qxh4 12. Bf4 $11 ) 7. Nc3 Nc6 { (This is presented as the only continuation but Nf6 looks at least as interesting.)} (7... Nf6 {(This is not discussed by JS.)} 8. Qd3 Nc6 9. Bf4 (9. Bg5 Be7 10. O-O-O h6 11. Bd2 a6 12. Nd5 Be6 13. Re1 $11 ) 9... Bg7 10. O-O-O Nh5 11. Be3 O-O 12. Kb1 Re8 13. Bf2 $11 ) 8. Be3 Nf6 {(Be7 and Nge7 have been tested in practice and are the continuations which JS covers. About the engine-recommendation which I trust more, is nothing written.)} 9. Bd3 Bg7 10. Qd2 Nb4 11. O-O-O Nxd3 12. Qxd3 O-O 13. h3 gxh3 $11 *

In the book a much simplified view is given of the possibilities so partly misguiding how complex the task is for both players. I do have to admit that the type of position seems very attractive for the true King's gambit adept as white has nice compensation for the sacrificed pawn.

Next I covered Bg7. My analysis completely overlap the book but I am less optimistic about whites chances in the mainline after replaying a recent key-game.

[Event "RUS/Emerald2011 (RUS)"] [Site "ICCF"] [Date "2011.12.15"] [White "Nepustil, Frantisek"] [Black "Fetisov, Aleksey Anatolievic"] [Result "1/2-1/2"] [WhiteElo "2390"] [BlackElo "2351"] [PlyCount "114"] 1. e4 e5 2. f4 exf4 3. Nf3 g5 4. Nc3 d6 5. d4 Bg7 6. g3 {(With a transposition we are in the line which JS recommends against the Fischer-defense with 5.g3 Bg7.)} Nc6 7. d5 Ne5 8. gxf4 gxf4 9. Bxf4 Bg4 10. Bb5 (10. Be2 Nxf3 11. Bxf3 Qf6 12. O-O {(Black has a very comfortable position.)}) 10... Kf8 11. Be2 Nxf3 12. Bxf3 Bxc3 13. bxc3 Qf6 14. O-O Qxf4 15. Bxg4 Qxe4 16. Bf5 Qe3 17. Kh1 {(Here JS stops and claims black has an extra pawn but his king is insecure.)} Nh6 18. Rf3 Qg5 19. Be6 Rg8 20. Qf1 Rg7 21. Re1 Kg8 22. Rg3 Qh4 23. Rh3 Qd8 {(Black avoids the repetition with Qg5 as the endgame is clearly better for black.)} 24. Rxh6 Qg5 25. Bxf7 Rxf7 26. Qg1 Rg7 27. Qxg5 Rxg5 28. Rg1 Rxg1 29. Kxg1 Re8 30. Rh4 b5 31. Kf2 Re5 32. Rd4 Kf7 33. a3 Rh5 34. h4 Kf6 35. c4 Ke5 36. Rg4 Rf5 37. Ke3 bxc4 38. Rxc4 Rf7 39. Ra4 a6 40. Rxa6 Kxd5 41. Ra4 Kc5 42. Rf4 Rg7 43. Kd2 Rg3 44. Rf7 c6 45. Rxh7 Rxa3 46. h5 Rh3 47. Ke2 Kd4 48. Rh6 Rh2 49. Kf3 Ke5 50. Rh8 Rxc2 51. h6 Rh2 52. h7 Kd5 53. Kg4 c5 54. Kg5 Kd4 55. Kg6 d5 56. Rg8 Kd3 57. h8=Q Rxh8 {(A clever draw but black clearly had all the fun. So I find JS somewhat too optimistic about whites chances in this line.)} 1/2-1/2

Is the author not familiar with this game as it is clear white had to fight hard for the half point or worse was this game on purpose ignored? In the book games of 2013 are used so weird. Finally I don't want to deny the reader of a very interesting possibility which not by coincidence also seems to give the best chances to refute the idea.

[Event "KG Fischersysteem 5.g3 h6"] [Date "2014"] [White "?"] [Black "?"] [Result "*"] [ECO "C34"] [PlyCount "29"] 1. e4 e5 2. f4 exf4 3. Nf3 d6 4. d4 g5 5. g3 h6 {(Probably the most critical test.)} 6. Nc3 fxg3 (6... Bd7 $146 {(An interesting novelty neither mentioned by JS after which it is very difficult for white to keep the balance.)} 7. Qd3 (7. Bc4 Nc6 8. O-O Na5 9. Be2 fxg3 10. hxg3 Bg7 11. d5 Ne7 $13 ) (7. Bd3 fxg3 8. hxg3 Bg7 9. Be3 Nc6 10. Be2 Qe7 11. Qd2 Nf6 12. d5 Ne5 13. O-O-O Nxf3 $13 ) 7... Nc6 8. a3 Bg7 9. gxf4 g4 10. Rg1 Kf8 11. d5 gxf3 12. dxc6 Qh4 13. Rg3 f2 14. Kxf2 $13 {(Houdini considers all this is just playable but I surely prefer blacks position.) }) 7. hxg3 {(More popular is h4 but equality I can not find for white.)} (7. h4 g4 8. Ng1 g2 9. Bxg2 Be7 10. h5 Bh4 11. Ke2 Nc6 {(Better than Bg5 from JS.)} 12. Nd5 Bf6 $15 ) 7... Bg7 8. Be3 {(The alternative Bc4 is also not easy playing.)} (8. Bc4 {(Again not covered by JS.)} Bg4 9. O-O (9. Rf1 Nf6 10. e5 dxe5 11. Bxf7 Ke7 12. Bg6 exd4 13. Qe2 Kf8 14. Ne4 Nc6 $13 ) 9... Nf6 10. Qd3 Nc6 11. e5 dxe5 12. Nxe5 O-O 13. Rxf6 Qxf6 14. Nxg4 Qxd4 $13 ) 8... Nf6 9. Qd2 {(A slight improvement on the important game Nigel Short - Luke Mc Shane played in 2011, London.)} Ng4 {(In Nightingale - Marczell, corr 2009 Nc6 was played with also a lot of fun for black but a same concept as chosen by Luke Mc Shane is here at least as interesting.)} 10. O-O-O c6 11. Kb1 Nd7 12. Bd3 Nxe3 13. Qxe3 Nf6 14. e5 Ng4 15. Qe2 $13 *

After the 8th move the author writes: "This is exactly the sort of position I want to reach with white in the King's Gambit. It's sharp, interesting and little explored with just five games from this position.... In such virgin territory it is impossible to give a comprehensive coverage." So the reader is left alone in the line which I consider as the most critical against 5.g3. Isn't it the task of the author to make a serious in depth analysis (as usually done in top-correspondence)? My idea of 6... , Bd7 is not covered and is as interesting as 6..., fxg3. Besides I also show several refinements and schemes which are useful for the 6..., fxg3 line.

Despite the remarks I do have to admit that the book is really good. It comprises a complete overview of the existing relevant theory with correct evaluations (although sometimes they are a bit too subjective). However I do understand too the comment of MNb on my previous article. Somebody possessing already a lot of material about the King's Gambit will find few or no new findings which have been worked out for weeks as I did for this article. The book is a reference for otb-players but a correspondence-player can just better consult the databases and make the research with an engine.

Brabo