Chess Opening Theory/1. d4/1...Nf6/2. c4/2...e5/3. dxe5/3...Ng4/4. Bf4/4...g5

The sideline 4...g5!? was not well regarded at the end of the 20th century. Borik wrote that "the move 4...g5 creates irreparable weaknesses in Black's camp", while Tseitlin decided "this extravagant tactical stroke weakens the kingside and, on general ground alone, cannot be good". Lalic warned that "Black should be aware of the risks he is taking by playing such a line". Nonetheless, the 4...g5 line has found new supporters in recent years, thanks to Black's wins in Van Wely – Mamedyarov, Ciudad Real 2004 (where White played 5.Bg3), and Graf – Asik, Kavala 2007 (where White played 5.Bd2).

The main reason why 4...g5 was not well considered is that this move weakens a lot of squares, mostly f5 and h5 (as they cannot be covered by the g-pawn anymore). White can try to exploit these weaknesses with the manoeuvres Bf4–d2–c3 (pressure along the diagonal a1–h8), Ng1–e2–g3–h5 (pressure against the squares f6 and g7) and h2–h4 (to open the h-file).

The retreat 5.Bg3
1.d4 Nf6 2.c4 e5 3.dxe5 Ng4 4.Bf4 g5 5.Bg3

For years, the reaction 5.Bg3 was not well considered because it does not make the most out of Black's provocative fourth move. Tseitlin considered that "the bishop is in danger of staying out of play for a long time". But later Lalic found that 5.Bg3 was "just as effective" as 5.Bd2. Black concentrates on getting his pawn back, while White tries to get an advantage from the weakening of the black kingside. After the possible 5...Bg7 6.Nf3 Nc6 7.Nc3 Ngxe5 8.Nxe5 Nxe5 9.e3 d6 White has tried several ideas: White has also tried to quickly open the h-file with 7.h4 Ngxe5 8.Nxe5 Nxe5 9.e3 but after 9...g4! Black succeeded in keeping the file closed.
 * After 10.h4 h6 Tseitlin considers that Black has a satisfactory game.
 * After the natural but inoffensive 10.Be2?! Be6 Tseitlin thinks Black already has equality.
 * White also tried 10.Rc1, to remove the latent positional threat Bxc3 (that would double White's queenside pawns), indirectly cover the c4-pawn, ease the c4–c5 push in a distant future notably if Black castles on the queenside). After 10...Be6 11.b3 h5!? 12.h4 Ng6 13.hxg5 Qxg5 Lalic assesses that "White must be fractionally better due to the potential weakness of Black's h-pawn".
 * Lalic considers the best try to be 10.c5!, sacrificing a pawn to weaken Black's control on the e5-square and expose the black king further. The game Fraschini – Fuentes (Cuba 1995) continued with 10...O-O 11.cxd6 cxd6 12.Be2 Qb6 13.Qd2 Be6 14.O-O and Black felt compelled to complicate things with 14...d5!? to avoid being slowly constricted on the d-file.

The retreat 5.Bd2
1.d4 Nf6 2.c4 e5 3.dxe5 Ng4 4.Bf4 g5 5.Bd2

The retreat 5.Bg3 is less considered than 5.Bd2 (in order to place the bishop on the wide-open diagonal a1–h8), after which "White can expect a safe advantage". Then according to Lalic delaying the recapture with 5...Bg7 6.Bc3 Nc6 7.e3 Ngxe5 is not correct as White can gain an advantage by 8.h4 or 8.Qh5, so the immediate 5...Nxe5 is better. For some time 6.Bc3 was well considered because Black had problems to deal with the various positional threats: However the correct way for Black was found in 5...Nxe5 6.Bc3 Qe7 7.e3 Rg8! 8.Nf3 Nbc6 9.Be2 d6 10.Nd4 Bd7 11.b4 g4 with good counterplay for black on the kingside.
 * After 6...Bg7 7.e3 White already threatens to win a piece with the advance f2–f4. After 7...g4 8.Ne2 d6 9.Nf4 h5 10.Qc2 Joseph Staker suggests 10...Qg5 but play can continue by 11.Nd2 Bf5 12.Qb3 b6 13.c5! O-O 14.cxd6 cxd6 15.h4 Qh6 16.g3 Nbc6 17.Bg2 Rac8 18.O-O and White stands better thanks to the weaknesses in h5 and d6. Black has also tried 10...Na6 but it did not solve his problems.
 * After 6...Qe7 7.e3 Bg7?! White unleashes 8.h4! and Black has a creepy choice between 8...g4 (that would give the excellent f4-square for a white knight) and 8...h6 9.hxg5 hxg5 10.Rxh8+ Bxh8 11.Qh5 with advantage for White.

White's efforts then switched to 6.Nf3 to open the e-file, something that Black cannot really avoid as 6...Bg7 7.Nxe5 Bxe5 8.Bc3 would leave an advantage to White. For example 8...Qe7 9.Bxe5 10.Qxe5 10.Nc3 d6 11.e3 and Black is at a loss for an equalising line, White's advantage consisting in his ability to install his knight on the strong d5-square and to attack the weakened Black's kingside with the advance h2–h4. It is better for Black to continue with 6...Nxf3+ 7.exf3 when both 7...h5? and 7...Bg7 would fail to 8.Qe2+, so Black must try 7...d6 8.Qe2+ Be6 instead.