Leonardis Variation: Nf6 – A Solid, Balanced Position

White's d3 on move two is the hallmark of the Leonardis Variation — a modest pawn that supports e4 without committing to the sharp open-game structures of 2.Nf3 or 2.f4. After Black answers 2...Nf6 and White develops 3.Nf3, both sides have a knight out and a solid pawn centre. Stockfish at depth 16 gives the position a centipawn evaluation of -14, or roughly -0.14 in pawn units. That means the position is almost perfectly equal with a marginal edge for Black. In 5,708,974 Lichess games White wins 48.1%, draws 4.3%, and Black wins 47.6% — a coin-flip with outcomes decided almost entirely by the quality of play from here.

The engine recommends 3...Nc6 as Black's strongest continuation. The suggested principal variation runs 3...Nc6 4.g3 d5 5.exd5, leading to an open central battle where Black's well-placed knight on c6 supports the d5 push and keeps the position lively. In practice, Nc6 is also Black's most popular choice by a large margin: it appears in 2,564,979 games and White scores only 46.6% against it — the lowest White score among all Black's replies. Players who choose Nc6 consistently outperform Black's overall result in this line, confirming the engine's verdict. If you are playing Black, Nc6 into d5 is both theoretically sound and practically rewarding.

Black has several other reasonable continuations. The second most popular is 3...Bc5, played in 1,516,258 games, where White scores 47.9% — close to the global average. The move 3...d6 (665,930 games, White 50.7%) is slightly worse for Black in practice; White's extra space from d3 and e4 works well when Black keeps the centre closed with d6. The counter-thrust 3...d5 (439,645 games, White 48.9%) is ambitious but playable. Less common are 3...h6 (119,995 games, White 48.9%) and 3...c6 (91,050 games, White 48.4%). These last two are solid but passive — Black is not creating immediate threats and simply waits.

The grounding data flags two moves as mistakes in this position: 3...Bc5 and 3...h6, each costing roughly 111 centipawns compared to the best move 3...Nc6. While 3...Bc5 appears active, placing the bishop on c5 before resolving the central tension allows White to steer toward formations where the bishop is less effective. The engine prefers Nc6 because it immediately supports a d5 break, putting concrete pressure on White's centre. Similarly, 3...h6 spends a tempo preventing a bishop pin when White has not yet indicated plans to pin the knight. Both moves are not losing — they are labelled mistakes only in the sense that they let White off the hook for the slight theoretical disadvantage that should come with playing d3 early. Prefer Nc6 when in doubt.

White's strategic goal in the Leonardis is modest: equalize, avoid early weaknesses, and reach a middlegame where patient piece coordination decides the outcome. The d3 pawn is a double-edged choice — it keeps the position solid but means White cannot claim central superiority the way an early d4 would. Black, on the other hand, should aim for active play. The data shows Black's best practical results come from Nc6 followed by d5, contesting the centre directly. Against 3...d6, which scores White 50.7%, Black is too passive. The lesson is clear: in the Leonardis, fighting for the centre with Nc6 and d5 gives Black the best practical chances, consistent with the engine's evaluation of -0.14 favouring Black ever so slightly.