When we start to play chess we learn the rules and then stick to them. Or so you might think. The reality is somewhat different: time and again illegal moves occur in practice, whether accidentally or on purpose. Let’s take a look at some of the craziest examples!

by GM Ilja Zaragatski

Awkward moments such as the one suffered by this lady also occur to seasoned grandmasters! It’s just that they don’t usually look so good.

7: Korchnoi-Karpov, World Championship, Game 21, Moscow 1974

Let’s start with one of the classics of chess history. In the 21st game of his 1974 World Championship match in Moscow, a strong novelty soon earned Viktor the Terrible a clearly winning position:

In the diagram position, however, the World Championship challenger was suddenly no longer sure about the castling rules and here – on his move – he was forced to find the arbiter, Belgian Grandmaster Alberic O’Kelly de Galway, and check whether castling short was a legal move. After getting a reply in the affirmative he returned to the board, castled short and Karpov was forced to resign after 18…Bxc4 19.f4. So an illegal move wasn't actually played in the game, but the second best player on the planet should really know that castling is only forbidden if an opponent’s piece controls a square that the king has to cross.

6: Zaragatski – NN, NRW Championship blitz tournament, 1997, Hiddenhausen

A totally meaningless fun game Another wonderful pearl of chess history is up next in place no. 6: a blitz game by yours truly at the tender age of 11. Back then I was considered a promising young talent and managed to win the U13 NRW Championship by two points. I then got into the following position in the accompanying blitz tournament. My opponent, whose name I unfortunately can’t remember, had seconds on his clock when he mistakenly thought my rook was on h7 and replied 1…Kd8. That was of course an illegal move, but I didn't feel like claiming it and instead mated my opponent with 2.Kd6!

Blitz games are of course a perfect breeding ground for illegal moves. After all, who among us in the last remaining seconds of a pawn race hasn’t pushed his pawn forwards a square and a half at a time in order to gain a decisive tempo or two!? Another tip: in the Ruy Lopez, for example, when castling short you can calmly use your thumb to play Re1 and your little finger for h3. That wins two tempi and practically never fails 

5: Grischuk – Avrukh, Saint Vincent 2005 and Vallejo – Gelfand, Leon 2010

A kind of illegal move that very few perceive as such can be seen in the following double example. The game between Grischuk and Avrukh (diagram on the left) ended in stalemate after 64…Rg5+ 65.hxg5, while Vallejo-Gelfand was drawn after 66…Bxg5 67.Bxg5. So what? Well, according to the FIDE regulations a game of chess ends as soon as neither side is capable of giving mate. From articles 1.3, 5.2b and 9.6 it’s clear that the game is over the moment there’s no legal combination of moves that would make a mate possible. That’s the case in both examples after Black’s move. Note that in the example on the right, for instance, there’s no conceivable position with king and the same-coloured bishops in which either side could be mated! Of course this is taking things to extremes, but strictly speaking a lot of games continue like this after they’ve already ended in a draw 

4: Steiner – Colle, Budapest 1926

Another widespread kind of illegal move can be seen in the following three examples, which share fourth place. I christen this motif… a “piece beam” ©

Example 1:

13. ♘c3 c6 At this point Steiner accidentally knocked over his king, but placed it right back on the square it had come from. Or at least that’s what both players thought. In actual fact after its brief trip to the floor the monarch landed back not on h1 but g1, which went on to play a significant role in the game. Play continued as if nothing had happened: 14. c5 ♘d5 15. ♕b3 e6 16. ♗xd5 exd5 17. ♗g5 f6 18. ♖ae1+ ♗e7 19. ♖xf6 gxf6 20. ♗xf6 0-0 21. ♖xe7 ♖xf6 22. ♖xd7  A strong combination by Steiner, at least with the king on g1. In the more correct position with the king in the corner – the way this game found itself in the databases – you couldn’t entirely dismiss Rf1 and mate at this point. But instead there followed ♗xd7 23. ♘e4 ♖f7 24. ♕g3+ ♖g7 25. ♘f6+ ♔h8 26. ♕e5 ♗h3 27. ♘h5  and Colle conceded defeat. 1-0

Something very similar happened to the players in

Wilder - Arkell, London 1989 (left) & Mason - Winawer, London 1883 (right)

In the first case the board was set up again in another place due to annoying sunlight. Unfortunately the h3-pawn fell victim to that transfer, as it found itself a square back on h2, again without the players realising. And so the game went on. Wilder later played Bf1-h3, which enable him to gain a slight advantage, but not enough to win.

The Mason-Winawer game, however, was adjourned in the position on the right, but on the day they played it out the knight accidentally ended up on d7. Oddly enough, that went unnoticed by the players, so Black could continue with the strong 1…Nc5, with the idea of Ne4+, and soon won the game. A somewhat ridiculous turn of events for the player with White!

3: Torre – Reti, Baden Baden 1925 and Kindermann – Korchnoi

Another well-known type of disallowed move takes third place in our top list. The motif here is illegal castling!

The legendary Richard Reti, whose turn it was to move here with Black, found the strong move “long castles”, after which his king had found a bulletproof spot on the queenside. Carlos Torre, however, had something to say about the matter, and reminded the man who gave his name to the Reti Opening (1.Nf3) that his rook had previously moved to b8 and then returned to a8. Reti corrected his mistake and played 2…Kf8, since he’d already touched his king; the game ended in a draw after 31 moves.

The example on the right once again features Viktor Korchnoi. In his encounter with GM Stefan Kindermann he’d previously played Rg8 and then again Rh8, and in this position he simply castled. The faux pas initially went unnoticed and both players got into time trouble. But when reconstructing the moves after the time control they sensed something was wrong. In a clear unwillingness to get into another time trouble drama Kindermann and Korchnoi ultimately agreed to a draw. Viktor the Terrible! No, Viktor the terribly forgetful!

2: Kholmov – Lutivkov, Dubna 1976

I particularly like this example; it’s only the fact that it happened in real life almost 40 years ago and not in the internet age on a computer that surprises me a little. In this pre-pre-move era after 10…Bd7-g4 11.h3 Grandmaster Ratmir Kholmov was expecting nothing other than 11…Bxf3 and was already thinking how to respond to it. When his opponent instead played 11…Bc8 Kholmov didn’t take a close enough look and played his previously decided upon 12.Qxf3. Oddly enough Lutikov objected and noted that it was forbidden to capture your own pieces. White had touched his knight while capturing, and therefore here instead played 12.Ne1… and ultimately won the game regardless. So I would simply have accepted 12.Qxf3 as if nothing had happened and then confidently gone on to convert the extra piece 

1: Azmaiparashvili – Ivanchuk, 1994 and Romanovsky – Kasparyan, 1938

The first and best place on our rating list goes to two examples of exceptional sneakiness and trickery from players with the black pieces. Sadly the swindles were ultimately found out.

In the first case – an encounter between world class grandmasters Azmaiparashvili and Ivanchuk – Black is simply losing; he can’t defend against the dark-squared threats to his king. In the diagram position, however, Chucky uncorked the move 33…Qa3-g4, when after 34.Qf6 Kf8 Black would actually have saved his skin. Sadly, however, the queen can’t swerve like that, and it’s only with a great deal of imagination that you can get from a3 to g4 in one step. Ivanchuk had to take his attempt back and resigned, since otherwise there was no salvation on the horizon.

Things went a little more tragically for the well-known study composer Genrikh Kasparyan. In this highly promising position he went for the brilliant combination 52…Qe1+ 53.Kh2 Rxh3 54.Bxh3 (54.Kxh3 Qh4#) 54…Nf3#... only to realise that the knight was actually pinned, and he had to resign as he was a rook down in a hopeless position.

Has a similar mishap ever happened to you? Let us know!

See also:

• The 7 worst "wrong" outcomes to a game
• 7 stalemates every chess player needs to know
• The 7 worst movie chess scenes