Checkmate with Two Knights: Why It's Almost Impossible (Troitsky Line Explained)
⚡ Quick Answer (TL;DR)
- Two knights CANNOT force checkmate against a lone king — it's a forced draw (tablebase-proven). Mating positions exist, but the defender can always avoid them.
- The paradox: give the defender a pawn, and suddenly mate becomes possible — the pawn destroys the stalemate defence.
- The Troitsky Line tells you when it works: pawn blockaded on or behind a4, b6, c5, d4, e4, f5, g6, h4 = win.
- Even then it can take up to 115 moves — the 50-move rule often saves the defender.
Every beginner learns that two bishops can force checkmate. So surely two knights — worth the same points — can do it too? Here's one of the strangest truths in chess: they can't. And the reason why, plus the bizarre exception, makes this the most fascinating of all basic endgames — complete with a tragic genius, famous grandmaster failures, and mates that take over a hundred moves.
📚 Table of Contents
1. Why Two Knights Can't Force Mate
First, the confusing part: two-knight checkmate positions exist. Here's one — perfectly legal, engine-verified:
The problem is unique to knights: a knight cannot lose a tempo. A rook, bishop or queen can make a "waiting move" along a line and keep controlling the same squares; a knight that moves always gives up every square it was guarding. To finish the mate above, White needs the knights to arrive with the timing exactly right — and with best defence, that timing can never be forced. Endgame tablebases (computer databases of perfect play) have confirmed it: every K+2N vs K position is a draw with correct defence.
2. The Stalemate Problem
Try to trap the king in the corner, and this is what you get instead:
This is the wall every attacker hits: to cage the king you must take away all its squares, but taking away all its squares without check is stalemate. The knights simply can't do the final "one more tempo" trick. (Stalemate ruins more wins than any rule in chess — see our weird chess rules guide.)
3. The Great Paradox: a Pawn Makes It Winnable
Now the twist that delights chess players: if the defender has a pawn, the win can become forced. Extra material makes the defence worse — because the pawn destroys the stalemate defence. Look:
The technique: one knight blockades the pawn (freezing it), while the king and the other knight herd the enemy king into a corner. Only at the very last moment does the blockading knight abandon the pawn and rush over to help deliver mate — the pawn's few free moves are exactly the tempi the attacker needed.
4. The Troitsky Line
But when exactly does this work? That question consumed Alexey Troitsky (1866–1942), the father of the modern endgame study, who spent years analysing it and published his findings in 1937. His answer is the famous Troitsky Line:
Modern engines checked Troitsky's century-old pencil-and-paper analysis square by square. Grandmaster John Nunn, after verifying it with tablebases, called it astonishingly accurate. One catch survives: some wins need up to 115 perfect moves — far beyond the 50-move rule — so even a "won" position can be a tournament draw.
📖 The tragic genius behind the line
Troitsky's story is one of chess's saddest. He worked as a forest ranger in remote Russia, composing immortal endgame studies far from the chess world — analysing, by hand, questions like whether four knights beat a queen. He died of starvation in 1942 during the Siege of Leningrad in World War II, and his final manuscript was destroyed in the war. His line, though, turned out to be near-perfect — confirmed by computers he never lived to see.
5. Grandmaster Stories: Failures & Triumphs
Does this ending actually appear in real games? Rarely — and when it does, it humbles the greats:
- Lilienthal's double heartbreak: Grandmaster Andor Lilienthal reached the winning two-knights-vs-pawn ending twice within six years — against Norman and against future world champion Smyslov — and failed to convert both times. If a world-class GM can't do it over the board, you know it's hard.
- The classic success: Znosko-Borovsky vs Seitz (1931) remains the celebrated demonstration of the winning technique done right — including the beautiful detail that the king, not the knights, must waste the tempo by walking in a small triangle.
- Nunn & Korchnoi (1980): the ending appeared in the post-mortem analysis of a Nunn–Korchnoi game at the Phillips & Drew tournament — and neither player knew whether it was a win. That mystery is what sent Nunn to the tablebases years later.
- The modern conversion: Sergey Karjakin successfully executed the full technique against Samuel Sevian — blockading the pawn with one knight, herding the king to the corner, and converting a position most players would have drawn.
♙ Knights Are Tricky — Train Them
Knight tactics decide thousands of real games. Sharpen yours with free mate-in-2 puzzles from real play.
Complete your basic-mates knowledge: the two bishops checkmate (which is forced!), the 6 essential checkmate patterns, the 10 most famous checkmates, and how the knight moves. Formal rules: FIDE Laws of Chess; deeper theory: two knights endgame on Wikipedia.
Frequently Asked Questions
No — it's a forced draw with correct defence. Mate positions exist but can't be forced: knights can't lose a tempo, so the defender always escapes or gets stalemated.
The pawn ruins the stalemate defence — when the king is caged, the defender must move the pawn, and those spare tempi let the knights finish the mate.
The rule by Alexey Troitsky: a black pawn securely blockaded on or behind a4, b6, c5, d4, e4, f5, g6, h4 means the two knights win — though possibly in up to 115 moves.
Rarely: Lilienthal failed twice, Znosko-Borovsky vs Seitz is the classic success, and Karjakin converted it against Sevian in the modern era.
Was this guide helpful?