The Dunst Opening is a chess opening in which White opens with the move:

This fairly uncommon opening may have more names than any other: it is also called the Heinrichsen Opening, Baltic Opening, Van Geet Opening, Sleipnir Opening, Kotrč's Opening, Meštrović Opening, Romanian Opening, Queen's Knight Attack, Queen's Knight Opening, Millard's Opening, Knight on the Left, and (in German) der Linksspringer.

The names Heinrichsen and Baltic derive from Lithuanian chess player Arved Heinrichsen (1879–1900). The opening was analyzed and played by the New York master Ted A. Dunst (1907–1985), giving the opening its most popular name in the United States. The Dutch International Master (IM) and correspondence grandmaster Dirk Daniel ("Dick D.") van Geet (1932–2012) frequently played 1.Nc3, so it is often called the Van Geet Opening in the Netherlands. The appellation Sleipnir seems to come from Germany. Sleipnir is Odin's (Wotan in German) magical eight-legged horse, and chess knights are horses with up to eight different possible moves each turn. Czech Jan Kotrč (1862–1943), editor and publisher of the magazine České Listy, said the opening was analyzed by English players. Zvonimir Meštrović (b. 1944) is a Slovenian IM who often adopts this opening. Tim Harding refers to it as the "Queen's Knight Attack" (Harding 1974:8). National Master Hugh Myers called it "Millard's Opening" after Henry Millard (1824–1891), a blind correspondence chess player who drew with the opening in a simultaneous exhibition against Joseph Henry Blackburne. Blackburne later played the opening himself against Josef Noa in the London 1883 international chess tournament (Myers 2002:24–25). The German FIDE Master Harald Keilhack in his 2005 book on the opening states that it has also been referred to as the Romanian Defense, and that he prefers the neutral appellation "Der Linksspringer" or, in English, "the Knight on the Left" (Keilhack 2005:7).

The opening move 1.Nc3 develops the knight to a good square where it attacks the central e4 and d5 squares. Although quite playable, 1.Nc3 is rarely seen; it is only the eighth most popular of the 20 possible first moves, behind 1.e4, 1.d4, 1.Nf3, 1.c4, 1.g3, 1.f4, and 1.b3. As of February 6, 2009, out of the over 500,000 games in ChessGames.com's database, only 644—about 1 out of every 780—begins with 1.Nc3. The third-ranking 1.Nf3 is 66 times as popular. Some very strong correspondence chess players employ 1.Nc3 frequently, and it is occasionally seen over the board.

The reasons for 1.Nc3's lack of popularity are that it does not stop Black from occupying the center (while 1.Nf3 prevents 1...e5, 1.Nc3 does not prevent 1...d5 because the d-pawn is guarded by the queen), and it blocks White's c-pawn from advancing to c3 or c4 (often desirable) without first moving the knight. In addition, after 1...d5, the knight's position is unstable because Black may attack it with ...d4. Although 1.Nc3 develops a piece to a good square (unlike 1.Na3 or 1.Nh3), and does not weaken White's position (unlike, e.g., 1.g4 or 1.f3) or waste time (unlike, e.g., 1.c3), the above-stated drawbacks make it an inferior way of attempting to exploit White's first-move advantage. Of the 644 games with 1.Nc3 in ChessGames.com's database, White won 34.8%, drew 23.9%, and lost 41.3%, for a total winning percentage of only 46.75%. White scores much better with the more popular 1.e4 (54.25%), 1.d4 (55.95%), 1.Nf3 (55.8%), 1.c4 (56.3%), and 1.g3 (55.8%). As of October 2018, however, for the 38,043 games with 1.Nc3 in Lichess.org's database, White has a total winning percentage of 57.25% (55% won, 5% drawn, 40% lost), although it is still only the tenth-most popular first move for White. As a testament to Nc3 as an interesting move choice to simplify opening theory yet being effective: After 1.e4 d5 the usual 2.exd5 is 48% win rate for White while 2.Nc3 has a 49% win rate as of 30 July 2022 – albeit with less sample size but still over 300,000 games. This could arise from the mainline of this article from 1.Nc3 d5 2.e4 which shows in fact the Dunst system is actually being regularly used online with great effect.

This is one of Black's best replies, occupying the center and undermining the unsettled position of White's knight. White can prevent 2...d4 by playing 2.d4 themselves, transposing to a well-known position in the Queen's Pawn Game, where after 2...Nf6 White can choose to play a Jobava London system with 3.Bf4, a Veresov Attack with 3.Bg5, or transpose to the Blackmar–Deimer Gambit with 3.e4?!.

Also possible is 2.Nf3 (and if 2...d4, 3.Ne4), a sort of Black Knights' Tango with an extra move (Harding 1974:10).

A third line is 2.e3, which Keilhack calls "the Müller game", when White anticipates 2...e5 (other moves are also playable) 3.Qh5!?, e.g. 3...Nc6 4.Bb5 Qd6 5.d4 exd4 6.exd4 Nf6 7.Qe5+! Be6?! (Keilhack recommends 7...Kd8!! 8.Bxc6 bxc6 9.Nf3 Bg4) 8.Bf4 0-0-0 9.Bxc6 Qxc6? 10.Nb5!, when White wins at least a pawn (Keilhack 2005:307–11).