The Hardest Chess Puzzles Ever Composed
Some chess puzzles transcend training. They're art. Composed studies — positions invented not from real games but as logical puzzles with hidden, counterintuitive solutions — push the limits of human calculation and intuition. These are the legendary puzzles that have stumped grandmasters for over a century.

Try the hard ones
Denser positions with more material and a forced mate in three. The defence has real resources here, so the first move often looks nothing like an attack.
What makes a chess puzzle "hard"?
Hardness in chess puzzles comes from multiple sources:
- Counterintuitive moves. The right answer violates standard chess principles — sacrificing the queen, retreating instead of attacking, underpromoting instead of queening.
- Deep calculation. The solution involves 10+ moves of forced sequences with multiple branches.
- Quiet moves. Sometimes the key move isn't a check or capture but a positional move whose purpose only becomes clear 6 moves later.
- Unusual themes. Stalemate tricks, threefold repetition, underpromotion to a knight or bishop instead of queen.
The most famous composed studies
Reti's classic king study
White: King h8, Pawn c6. Black: King a6, Pawn h5. White to play and draw.
The position looks lost — Black's pawn will promote, White's king is too far away. But Reti's brilliant solution shows that the king can do two jobs at once: race the pawn AND support the white pawn's promotion. 1.Kg7! threatens both, and Black cannot stop both threats. The study revolutionized how chess players think about king geometry.
The Saavedra position
White: King b6, Pawn c7. Black: King a1, Rook d5. White to play and win.
White looks like they'll only draw — promoting to a queen allows Rd4+ and Rxc4 with a drawn K+P vs K+R. The brilliant trick: underpromote to a rook (not queen!) which avoids the skewer. After 1.c8=R! Ra5! 2.Kb6! threatens both Kxa5 and Rc1#. Black cannot prevent both — White wins.
The Saavedra position is famous for showing that the lowly rook can beat the queen in specific positions, because the queen creates a stalemate trap the rook avoids.
The Troitzky line
Troitzky composed dozens of studies showing K+N+N vs K+P is winning under specific conditions. The complete proof — when two knights can checkmate a king with a pawn — defines the "Troitzky line" in endgame theory. A theoretical curiosity that's also a practical winning technique in rare endgames.
The Excelsior problem
Loyd's most famous mate in 5 problem. The thematic point: a pawn promotes after a long march, transforming during the solution into the mating piece. Considered one of the most elegant composed mates ever.
Rinck's queen vs rook studies
Rinck composed thousands of endgame studies. His Q vs R studies showed brilliantly how the apparently overwhelming queen can be neutralized by precise rook play in specific geometric setups. Many serve as templates for real-world Q vs R endgames.
Why study composed problems?
Composed studies seem impractical — these positions rarely occur in real games. But the training benefit is real:
- Counterintuitive thinking. Studies force you to consider moves that violate normal principles. This expands your search in real games.
- Calculation depth. No tactical exercise builds depth like a 10-move composed study.
- Aesthetic appreciation. The beauty of chess composition inspires deeper love for the game.
- Pattern enrichment. Underpromotion, stalemate tricks, and geometric oddities expand your tactical vocabulary.
The hardest game positions ever played
Beyond composed studies, certain real-game positions are famous for their difficulty:
- Karpov-Kasparov 1985, game 16 — Kasparov's strategic masterpiece. Beautiful to study but extremely difficult to understand fully.
- Kasparov-Topalov 1999 — the modern Immortal, featuring a king walk to a5 across the board. Calculating the complete combination is grandmaster work.
- Polugaevsky-Nezhmetdinov 1958 — Nezhmetdinov's queen sacrifice and 13-move conversion. Studied for decades.
How to approach hard puzzles
- Set the position up on a real board. Pixels make hard puzzles harder. Tactile pieces let you focus on patterns.
- Take your time. 30 minutes on one hard puzzle is more valuable than 30 easy puzzles in the same time.
- Don't peek at the solution. Even if you can't solve it, the struggle teaches calculation.
- When you give up, study the solution slowly. Understand why each move works. Try to articulate the underlying principle.
- Replay the puzzle a week later. See if the pattern stuck.
Computer engines and hard puzzles
Modern engines solve essentially every composed study in seconds. This has fundamentally changed the field — composers now use engines to verify solutions, and many "hardest" historical studies have been found to have alternative solutions the original composer missed.
For training, use engines AFTER attempting the puzzle. Engines tell you the answer instantly — use them as verification, not as a shortcut past the work.
Modern hard puzzle sources
- EG (Endgame Studies magazine) — quarterly journal of new composed studies.
- Yet Another Chess Problem Database — online searchable database of composed problems.
- Chess.com Daily Problem — often features composed studies and hard positions.
- Lichess Studies — community-curated collections of legendary positions.
Questions about these puzzles
Once these feel comfortable, the natural next step is a full game — play a game of chess online at whichever level gives you a real contest.