1.Nf3 Nf6 2.c4 g6 3.Nc3 d5 4.cxd5 Nxd5 5.Qa4+ "Don't bring out your queen early, especially if it's just for check", is a piece of instruction that we routinely teach amateurs. As I noted in the analysis to the Van D Mortel-Stripunsky, though, grandmasters not only know the rules, they know when to break them. Here Rohde has a very specific and concrete purpose in mind: he is looking to exploit the extremely temporary weakness of Black's diagonal to his rook, before Black comfortably fianchettoes his bishop there, achieving a harmonious set-up. These are the nuances and details that characterize contemporary gm play.
5...c6 6.Qd4 Precisely. Now Black has the unenviable choice of retreating his knight, allowing a queen swop which forfeits the right to castle, and the game continuation, which is unambiguously an unfavorable alteration of Black's pawn structure. His f-pawn will block his own bishop, and the light-squares around the king are weakened.
6...f6 7.e4 e5 Now Black has a way to either avoid the queen trade, or extract his own concession. If White wants to insist on a queen trade, then he can play 8.Qd3 (see variation below)
8.Nxe5 It appears that things are getting out of control, but in fact we are still within the bosom of theory. Of course, Black cannot hastily capture White's knight, as the queen recapture with check targets Black's rook in the corner. Nevertheless, he has a path to reestablishing material equality. [8.Qd3 Nb4 9.Qxd8+ Kxd8 but now there is no comfortable way to protect the c2 square, so he may need to forfeit castling himself with Kd1.]
8...Nxc3 9.Qxc3 Qe7 Here it is. Black will regain his pawn.
10.Nc4 Qxe4+ 11.Ne3 Be7? A sign that Paragua is on his own. Passive moves are rarely strongest, especially in quick chess. One of my favorite players and a mighty theoretician, Croatian GM Zdenko Kozul here played far more actively with 11...Bb4!, sacrificing his f-pawn for activity. After 12.Qxf6 Rf8, Black gains tempi against White's queen to activate his forces. With so many pieces bearing down on White's position, the pressure is palpable, and Kozul went on to win in Wirthensohn-Kozul 1991. After Paragua's lackluster move, White takes control of the position. [11...Bb4 ]
12.Bc4 Remember that earlier move, 6...f6? I said it weakened the light squares. Now we see the impact of it - Black cannot castle, and White's light-square bishop has a monster diagonal into Black's heartland. All of this was a result of the gm queen maneuver beginning with the strange-looking 5th move, 5.Qa4+. These are precisely the kinds of games that produce pride: it's what's meant when players speak of following a logical thread throughout the game, and can be one of a top player's greatest aesthetic joys from chess.
12...Nd7 13.d4?! Rohde was likely concerned about Black's knight coming to e5, an optically fantastic spot and a disturbance to White's bishop. Nevertheless, the principled 13.0-0, taking a huge lead in development, was the way to go. The proof lies in a very subtle point, one which is nearly impossible to spot in quick chess: [13.0-0 Ne5 14.Re1! Nxc4 15.Nc2!! White sacrifices a piece, but now his lead in development leads to a winning advantage. White has so many threats: to Black's queen, to the f6 pawn, to Black's knight. He cannot cover them all.]
13...Nb6 14.Bd3 Black has succeeded in driving White's bishop from the commanding diagonal. Just like that, the game is even again.
14...Qh4? A strange move at first glance, objectively a dubious one, and in this game, ultimately the cause of Black's demise. Black's idea is to retain pressure on White's d-pawn, as he anticipates castling queenside. True, White's d-pawn is isolated, but the queen can't be retained on that rank, and White will extract an advantage from this decentralization. The main issue is that Black's queen is short on squares over there. 14...Qe6 was the sound alternative.
15.0-0 Be6 16.Re1 [16.g3 Now is a perfect moment for this, taking Black's queen away from its target. White has an advantage after 16...Qh3 17.Re1 0-0-0?? 18.Qa5 Kb8 19.Bf1 when the weakness of Black's queen shows. Black is short on squares, a result of his decentralization, and his bishop pile-up on the e-file, opposite White's rook, is ominous. White wins.]
16...0-0-0 17.Nc2 Now, by contrast, White feels obliged to go passive.
17...Nd5 18.Qa5 Bf5?? The move looks very natural, since the e7 bishop is now defended, but White finds a tactical refutation, based on the weakness of Black's king.
19.g3! Qg4 20.Rxe7! Qd1+?? Paragua loses his cool. For now, the c7 square is defended by Black's knight, so he should just recoup his material with 20...Bxd3. The frightening-looking 21.Bf4 can be met with 21...Rd7, and Black has not lost control of the c7 square. [20...Nxe7 This loses, though 21.Bf4 Nd5 22.Bxf5+ gxf5 23.Ne3 Qg7 24.Nxd5 Rxd5 25.Qxa7 Black managed to retain eyes on c7, but the attack crashes through all the same. White will decimate Black's pawn defense and send his king into the wilderness. A resignable position for Black.]
21.Ne1? It was important for White to preserve his light-square bishop, as it can imminently become a crucial attacker: [21.Bf1 Qxc2 22.Bf4 Rd7 23.Rc1 Not the best move, but the most direct illustration 23...Qxb2 24.Rxc6+ bxc6 25.Ba6+ and the cost of this attack to Black becomes clear.]
21...Rd7?? Amazingly, Black can survive after [21...Bxd3! 22.Bf4 Qxa1 23.Rc7+ Kb8 and White has nothing better than perpetual check 24.Rd7+ Kc8 25.Rc7+ etc. Now, by contrast, White wins prosaically. Black never recouped his piece deficit, so Rohde cashes in simply.]
22.Bxf5 gxf5 23.Rxd7 Kxd7 24.Qd2 Qh5 25.Qd3 Re8 26.Bd2 f4 27.Nf3 Rg8 28.Re1 Kc8 29.Qe4 Kb8 30.Bxf4+ Kc8 31.Qe6+ A magnificent display of fireworks! 1-0