Black's first move 1...Nc6 meets two very different games, and the opponent's rating decides which one. Against 2.d4, the Nimzowitsch Defense has to prove itself: our engine puts the position at +0.65 for White. Against 2.Nf3, the likeliest move for players up to 1800, Black can simply play 2...e5 and reach an ordinary Open Game. This page gives the numbers behind that split, line by line, from both sides of the board.
The Nimzowitsch Defense is 1.e4 Nc6, ECO B00. The page is called the Nimzowitsch Defence on Wikipedia, which also calls it somewhat uncommon and hypermodern: Black does not fight for the centre with a pawn, but lets White build one and then attacks it. Do not confuse it with the Nimzo-Indian Defence or the Nimzo-Larsen Attack. Aron Nimzowitsch's name appears on all three, but they are different openings.
We ran the maia3-5m checkpoint (see how Maia chess actually works) on the position after 1.e4 Nc6 at seven rating bands from 1000 to 2200 and graded every legal move with Stockfish 18. The percentages are the model's predicted move probabilities for a player at that rating. They are not database frequencies and not win rates. Evaluations are from White's side, so +0.65 means White is better by 0.65 pawns, and "Black +0.33" means the engine prefers Black by 0.33 pawns. Line checks are separate local Stockfish runs at 10 seconds a position, reaching depths of 21 to 25. Treat any gap under about 0.15 pawns as a tie.
Where the name comes from
The usual line is that Nimzowitsch invented it. The history says otherwise. Edward Winter's Chess Notes, citing Myers (1995), reports that 1.e4 Nc6 was known at least as far back as the sixteenth century and that several players used it in master play in the 1800s, before Nimzowitsch was born in 1886. In Germany it is often called the Fischer Defence after Eduard Fischer (1831 to 1897), mostly a correspondence player, and Winter adds that Wilfried Paulsen learned it from Fischer in 1858 and 1859. Winter's page quotes Modern Chess Openings (11th edition, 1972) calling it a bizarre defence first advocated by Fischer, which Nimzowitsch studied deeply in 1919 and concluded was sound.
Nimzowitsch played it against Duras at Ostend in 1907 and annotated a game with it in 1920. Myers says he was the first to play it with frequent success in major international competition, and that is why the name stuck. So read the name as "popularised by" and not "invented by". For the spelling, Winter lists Nimzovich, Nimzowitsch, Niemzowitsch and several others as found in print, and the older Chess.com pages use Nimzovich. This page uses the Lichess spelling, Defense, and treats the rest as the same opening.
What players actually play against 1...Nc6
As of 29 October 2026, the Lichess page for 1.e4 Nc6 lists White's second moves by share of games, with a rating filter of 1000 to 2500 and all speeds when we fetched it: 2.Nf3 at 41 percent, 2.d4 at 29, 2.Nc3 at 9, 2.Bc4 at 8 and 2.f4 at 4. The page gives each row two unlabelled percentages and no game counts, so we use the shares only. Our model is a different instrument: it predicts what a player at each rating would choose, rather than counting games. The two broadly agree on the order, and ours adds the rating curve.
| Rating | 1000 | 1200 | 1400 | 1600 | 1800 | 2000 | 2200 |
|---|---|---|---|---|---|---|---|
| 2.d4 (+0.65) | 17.5% | 19.6% | 21.2% | 24.5% | 31.4% | 42.0% | 48.5% |
| 2.Nf3 (+0.31) | 39.5% | 44.4% | 45.7% | 44.5% | 40.7% | 35.6% | 34.3% |
| 2.Nc3 (+0.18) | 12.1% | 9.0% | 7.6% | 7.1% | 7.0% | 7.3% | 8.9% |
| 2.Bc4 (Black +0.33) | 11.6% | 11.4% | 11.1% | 9.7% | 7.1% | 4.1% | 1.8% |
| 2.f4 (Black +0.31) | not top five | not top five | 3.4% | 4.9% | 6.5% | 5.9% | 2.8% |
Three things stand out. First, the engine's top move is not the human favourite for most of the range. 2.d4 is best at +0.65, but 2.Nf3 (+0.31) is the likeliest move at every band from 1000 through 1800, between 39.5 percent and 45.7 percent. That is a gap of 0.34 pawns between the best move and the one players pick. Second, 2.d4 climbs steadily, from 17.5 percent at 1000 to 31.4 percent at 1800, then jumps to 42.0 percent at 2000 and 48.5 percent at 2200, which is where it overtakes. At those bands the likeliest move and the best move are the same.
Third, 2.Bc4 is the second or third most likely move from 1000 through 1800, and sits at 9.7 to 11.6 percent from 1000 through 1600, yet the engine puts it 0.33 in Black's favour, the worst of the five rows. 2.f4 is the same story in miniature at Black +0.31. Neither is a disaster, but both hand Black a small edge.
To be fair to 2.Nf3, it is a good move. It keeps +0.31, a normal opening edge, and it does not walk into anything. Wikipedia says some databases show it as the most common reply, and quotes Jeremy Silman that most players, even grandmasters, answer 1...Nc6 with 2.Nf3 to avoid prepared lines. ChessMood recommends it as well. The point of the data is narrower: below 2000, most players let Black steer into an ordinary 1.e4 e5 game, and it is 2.d4 that actually tests the Nimzowitsch.
2.d4: the line that tests it
After 2.d4, Black has three real choices, and the engine ranks them in a clear order: 2...d5 at +0.60, 2...e5 at +0.75 and 2...d6 at +0.90. Black cannot get below +0.60 in any of them. This is the cost of the opening in one line: our King's Pawn checks put 1...e5 at +0.36, so against 2.d4 Black pays roughly a quarter of a pawn for the unusual first move.
2.d4 d5: the Scandinavian Variation
Wikipedia calls 2...d5 the Scandinavian Variation, and three White moves matter. All three are checked below.
- 3.e5. Black's best is 3...Bf5 at +0.52, then 3...f6 at +0.57 and 3...a6 at +0.65. After 3...Bf5 4.c3 e6, White's best is 5.Be2 at +0.50. Wikibooks calls this a superior French structure without the passive bishop, and ChessMood calls it strange and French-like. If the plans are familiar from the French Defense, that is the right comparison, though Black has lost the usual ...c5 break. White's edge is half a pawn, which is real but not scary.
- 3.exd5. After 3...Qxd5 4.Nf3, Black's best reply is 4...e5 at +0.32, with 4...Bg4 at +0.35. ChessMood calls the 4...Bg4 5.Be2 O-O-O setup very active and pleasant for Black. The engine's number is milder: White's best 6.Be3 is +0.27 there. The edge is small, so ChessMood's verdict holds up. Compare the plans in our Scandinavian Defense guide, because the structure is the same with a knight misplaced on c6.
- 3.Nc3. Here the sources disagree. ChessMood says 3...dxe4 4.d5 Ne5 is very playable for Black, especially at a lower level. The ChessBase review has Andrew Martin answering 3.Nc3 with 3...e6 instead, the cautious choice. The engine sides with ChessMood. After 3.Nc3, Black's best is the capture 3...dxe4 at +0.25, while 3...e5 is +0.70 and 3...e6 is +0.72. After 4.d5 Ne5, White's best 5.Bf4 gives +0.50, and ChessMood's 5.Qd4 gives +0.31. That is an ordinary edge, not a danger. By these numbers the "safe" 3...e6 is the worse choice, by nearly half a pawn.
2.d4 e5: the Kennedy Variation
Wikipedia names 2...e5 the Kennedy Variation and says Tony Miles favoured it. The tidy move is 3.Nf3, which transposes to a Scotch Game structure, and the sharper try is 3.d5 Nce7 and 4...Ng6, the Black Knight's Tango. We did not check 3.d5. After the natural 3.dxe5 Nxe5, White's best is 4.Nc3 at +0.66, with 4.a3 at +0.61 and 4.Nf3 at +0.56, which fits Wikibooks marking 4.Nf3 as a slight White edge. White keeps about two thirds of a pawn.
The other 2.d4 replies
2...e6 leads to a French where Black cannot play ...c5, and the engine confirms it: White's best 3.Nf3 is +0.92. 2...d6 is the Mikenas Variation, which Wikibooks says concedes the initiative to White, and 3.d5 reaches +0.86 in our check. Both are worse than 2...d5 by 0.26 to 0.32 pawns, which is the point of the Mikenas warning: do not choose them because they look solid.
2.Nf3: the move most players choose
After 2.Nf3 the engine's three leading Black moves are 2...e5 at +0.23, 2...d6 at +0.62 and 2...d5 at +0.68. Black's best reply is 0.39 pawns better than the next one, which is the clearest honest statement about the opening: against 2.Nf3, 2...e5 is the move, and every other reply is worth noticeably less.
2.Nf3 e5: a transposition
2...e5 reaches the Open Game, the same position as 1.e4 e5 2.Nf3 Nc6. Wikipedia lists it as a transposition to a double king's pawn opening, and Reinke's ChessBase review names 2...e5 and 2...d6 as Black's two valid answers. Our late-position data shows White's best at +0.26 there. The King's Pawn Opening page covers this exact position in a band table, so we do not repeat it. Maia agrees this is Black's habit: at 1400, the likeliest Black reply to 2.Nf3 is 2...e5, at 43.5 percent. If your aim is simply to avoid theory while ending up in a main line, this is the safe way to use 1...Nc6. It also means the Nimzowitsch Defense, against many opponents, is just another route into the Italian, Ruy Lopez or Scotch.
2.Nf3 d6: the Williams Variation
The standard continuation is 3.d4 Nf6 4.Nc3, which gives a Pirc-like structure (see the Pirc Defense), and Wikipedia calls 2...d6 a solid but less dynamic option. Black's active try 4...Bg4 is where the opinions split. ChessMood says 5.Be3 almost refutes all of Black's ideas, ChessBase's Reinke says the line is considered better for White but still playable, and Wikibooks marks 5.Be3 e6 6.h3 a slight White edge. The engine gives White's best, 5.Be3, +0.94 at depth 23, with 5.Bb5 at +0.80 and 5.Be2 at +0.75. Even the weakest of those is three quarters of a pawn. So "almost refuting" is strong language, but the number is the largest edge in any main line we checked except the Colorado Gambit, and Reinke's "playable" still fits a player who accepts being worse.
Wikipedia also lists the Keene Attack, which comes from a different move order, 3.d4 Bg4 4.d5 Ne5 5.Nxe5, and credits it to Raymond Keene's 1964 game against E. Fielder. We did not check it.
2.Nf3 f5: the Colorado Gambit
Wikipedia calls 2...f5 somewhat dubious. ChessMood says very dubious. White's best capture 3.exf5 gives +1.06 at depth 23, and even 3.e5 is +0.66. Take the pawn and play for development. It is a pawn-sized error, not a forced loss.
2.Nf3 d5 and the rest
2...d5 is the engine's third choice at +0.68 and a delayed Scandinavian: after 3.exd5 Qxd5 4.Nc3, ChessMood says it is the Scandinavian with a misplaced knight on c6. Reinke goes further and says that after 4...Qa5 5.Bb5 Bd7 6.d4 White is clearly better in all lines, and the engine puts 2...d5 0.45 pawns behind 2...e5. We did not check the lines after 4.Nc3.
2.Nc3 and the fringe
2.Nc3 is the third most likely move at 1000 (12.1 percent) and sits near 7 percent for most of the range. The engine ranks it third at +0.18, behind 2.d4 and 2.Nf3. Black's best reply is 2...e5 at +0.16, nearly level, with 2...Nd4 at +0.62 and 2...d6 at +0.74. Master In Chess names 2...e6 the French Connection, but we did not check it, and 2...e5 is the engine's choice. For White, 2.Nc3 is a fine way to reach a quiet middlegame, and a poor way to punish the opening.
The remaining moves are rarer. 2.Bb5 is the Pseudo-Spanish Attack and 2.b4 the Wheeler Gambit, per Master In Chess. Our early position puts Bb5 at Black +0.02, essentially level, so a Ruy Lopez-looking move does not gain anything here.
If you play it, and if you face it
Our data has no results by rating, so this is a guide to what the model and the sources suggest, not a score table.
- As Black under 2000. Most opponents play 2.Nf3, and 2.d4 is second at every band. Answer 2.Nf3 with 2...e5 and you are in a normal Open Game. Learn the French-like 3.e5 Bf5 plan and take on e4 after 3.Nc3.
- As Black at 2000 and above. Expect 2.d4 about four times in ten at 2000 and nearly half the time at 2200. ChessMood suggests keeping the defence as a reserve weapon against 2400 and stronger players, and it notes top players use it mostly in blitz and rapid. The engine says a prepared 2.d4 gives White about +0.6 or +0.7.
- As White. 2.d4 scores best, and 2.Nf3 is simpler. If you want a quiet game with +0.31, play 2.Nf3 and see whether Black finds 2...e5. If you want pressure, play 2.d4 d5 3.e5, which is French-like, or 3.exd5 for the Scandinavian-with-a-knight-on-c6 plans.
Is it sound?
The sources split. Kasparov and Keene, quoted on Wikipedia and Master In Chess, say the opening has never been fully accepted as dependable but is sound, and offers the maverick plenty of foreign territory. ChessBase's own blurb calls it a fully valid line. ChessMood calls it very playable and practical, especially at a lower level, but says it does not solve all of Black's opening problems.
Our numbers fit "sound but a little worse". The position after 1.e4 Nc6 2.d4 is +0.65, no Black reply gets below +0.60, and the ordinary lines Black can choose sit between +0.23 and +0.94. Here is where it is objectively worse than 1...e5. Our King's Pawn checks put 1...e5 at +0.36, while after 1.e4 Nc6 2.d4 Black's best reply still leaves +0.60. After 2.Nf3, 2...e5 at +0.23 is fine, but 2...d6 is already +0.62. The opening looks fine in practice below 2000 largely because 2.Nf3 e5 is not really the Nimzowitsch any more. If you want a guaranteed comfortable game against every second move, 1...e5 is the safer choice (our King's Pawn Opening page has the scores). 1...Nc6 suits a player who wants a game off the main path and accepts a small price against 2.d4.
Where other resources are better
This page covers the position after 1.e4 Nc6 and the move that follows. The late position, after 2.Nf3 e5, is the Open Game, and our data stops at 2...e5. The line checks are single 10 second runs, not a repertoire book, and Maia3 5m is a small model of human choice, so its percentages are predictions. Wikipedia is better for the list of named variations and practitioners. The Lichess explorer has real game counts by rating and speed, which we have not reproduced, and as of 29 October 2026 its page for the opening carries no description. ChessMood is a good free walkthrough of White's replies with move orders, though it is written to sell a course and gives verdicts without engine numbers. The ChessBase review of Andrew Martin's DVD is useful for a Black repertoire. Winter's Chess Notes is the place for the history. None of the pages we read pairs each White try with an engine score and the rating curve, and that is the gap this one fills.
What to take from it
The Nimzowitsch Defense is 1.e4 Nc6, named after a player who popularised it, not invented it. After 1.e4 Nc6, 2.d4 is the engine's best move at +0.65 and 2.Nf3 the likeliest human move from 1000 through 1800, at 39.5 to 45.7 percent, while 2.d4 climbs from 17.5 percent at 1000 to 48.5 percent at 2200 and takes over at 2000. Against 2.Nf3, 2...e5 at +0.23 is clearly the best reply and transposes to an Open Game, and the other replies cost 0.39 pawns or more. Where sources disagree, the engine says 3.Nc3 dxe4 4.d5 Ne5 is an ordinary +0.50, the 2.Nf3 d6 3.d4 Nf6 4.Nc3 Bg4 line is +0.94 after 5.Be3, and the Colorado Gambit is +1.06 after 3.exf5. For how these openings change with strength, see chess openings by rating.
Frequently asked
- What is the Nimzowitsch Defense?
- It is 1.e4 Nc6, ECO B00. Wikipedia calls it a somewhat uncommon, hypermodern opening: Black lets White build a pawn centre, then restrains and undermines it. It is not the Nimzo-Indian Defence (1.d4 Nf6 2.c4 e6 3.Nc3 Bb4), which is a separate opening named after the same player.
- Is it spelled Nimzowitsch, Nimzovich or Defence?
- All appear. Edward Winter's Chess Notes lists Nimzovich, Nimzowitsch, Niemzowitsch, Niemtsowitsch, Nimzovitsch and Nimzovitch among the spellings found in sources. Wikipedia titles its page Nimzowitsch Defence, while Lichess writes Nimzowitsch Defense. They all mean 1.e4 Nc6.
- Is the Nimzowitsch Defense good?
- It is playable, not fully trusted. Kasparov and Keene, as quoted on Wikipedia, say it has never been fully accepted as dependable but is sound. In our Stockfish checks, White's best second move 2.d4 scores +0.65, so Black starts a little worse than after 1...e5, but no reply loses by force. Against 2.Nf3, 2...e5 holds at +0.23.
- What is the best reply to the Nimzowitsch Defense?
- 2.d4, at +0.65 in our data, is the engine's top move, with 2.Nf3 second at +0.31 and 2.Nc3 third at +0.18. Wikipedia's page quotes Jeremy Silman that most players, even grandmasters, avoid prepared lines with 2.Nf3, and ChessMood recommends it too. Both are sound, but only 2.d4 tests the opening itself.
- Who invented the Nimzowitsch Defense?
- Not Nimzowitsch. Winter, citing Myers (1995), says 1.e4 Nc6 was known at least as far back as the sixteenth century and was used in master play in the 1800s before Nimzowitsch was born in 1886. In Germany it is often called the Fischer Defence after Eduard Fischer. Nimzowitsch gets the name because he played it with frequent success in major events.
- Is the Colorado Gambit (2.Nf3 f5) any good?
- No. Wikipedia calls it somewhat dubious and ChessMood very dubious. Our local Stockfish check agrees: after 2.Nf3 f5, 3.exf5 leaves White ahead by 1.06 pawns. It is a pawn-sized mistake, not a refutation, so it is a poor choice against anyone who knows the capture.
Want to meet 1.e4 Nc6 against a model that answers like your rating?
Chessdrive runs the real maia3-5m checkpoint on-device as your opponent at any rating you set from 600 to 2600, so you can play the Nimzowitsch from either side and see what a player at your band actually chooses, then check where the game turned with a separate Stockfish review. Available now on iOS.
Keep reading
Updated