It takes two to "Tango"

My opponent in the following game found a new way to lose a pawn as White against the Black Knights' Tango (1.d4 Nf6 2.c4 Nc6!?). He began with 3.d5!?, which IM Richard Palliser calls "The Lunge" in his book on the opening, tango! At first blush, the move looks strong. White hopes to chase Black's knights with his pawns and force them back in disarray, à la Borochow-Fine, Kujoth-Fashingbauer, Marshall-Rogosin, and various games in the Halloween Gambit (1.e4 e5 2.Nf3 Nc6 3.Nc3 Nf6 4.Nxe5?! Nxe5 5.d4). Something like that did happen in Tate-Orlov, 1995. But if Black plays correctly, White gets little or no advantage with 3.d5. 3.Nf3 and 3.Nc3 are much more commonly seen, and more successful in practice.

White's 4.Qc2 was not in the spirit of The Lunge. White should continue his aggression with 4.e4 (when 4...Nxe4?? 5.Qd4 wins for White) or even Tate's 4.f4!?, offering the c-pawn as a gambit (4...Nxc4 5.e4). White's 7.a3? was an unfortunate theoretical novelty. Another passive, time-wasting move, it dropped a pawn to the simple tactic 7...Bxd2+ 8.Nxd2 exd5 9.cxd5 Nxd5! The lost pawn should have been the extent of the damage. Surely no one would be so naïve as to take the knight, which would give me the choice of two mates in two, or a mate in four? To my astonishment, my opponent fell into it hook, line, and sinker. I'm now 110-0-0 on GameKnot, with an Orwellian 1984 rating.