Study of the month: An impossible move

by Siegfried Hornecker
2/4/2017 – In our previous monthly study column we saw how powerful pawns are. The subject is amplified this time with three positions, two from games and one a study by two prolific chess composers from Georgia. In that study White would like to move his pawn backwards, which of course the rules of chess do not allow. So he has to promote it, very carefully, and then move the queen or rook to the square the pawn could not occupy. A remarkable idea and an amazing study.

ChessBase 17 - Mega package - Edition 2024 ChessBase 17 - Mega package - Edition 2024

It is the program of choice for anyone who loves the game and wants to know more about it. Start your personal success story with ChessBase and enjoy the game even more.


An impossible move

By Siegfried Hornecker

We have seen in the first installment of this column how powerful pawns are. Indeed the value of a pawn is said to increase considerably, already when it reaches the sixth rank. If several pawns work together, miracles can happen that would otherwise not be possible, such as in two game fragments, one more famous and the other less, but both astonishing.

You probably know that you can move pieces on our replay boards to analyse, and even start an engine to help you. You can maximize the replayer, auto-play, flip the board and even change the piece style in the bar below the board. At the bottom of the notation window on the right there are buttons for editing (delete, promote, cut lines, unannotate, undo, redo) save, play out the position against Fritz and even embed our JavaScript replayer on your web site or blog. Hovering the mouse over any button will show you its function.

[Event "Castilian Championship (?), Madrid"] [Site "?"] [Date "1933.??.??"] [Round "?"] [White "Ortueta Esteban, Martin"] [Black "Sanz Aguado, José"] [Result "0-1"] [ECO "C00"] [Annotator "Hornecker,Siegfried"] [PlyCount "70"] [EventDate "1933.??.??"] 1. e4 e6 2. d3 d5 3. Nc3 Nf6 4. e5 Nfd7 5. f4 Bb4 6. Bd2 O-O 7. Nf3 f6 8. d4 c5 9. Nb5 fxe5 10. dxe5 Rxf4 11. c3 Re4+ 12. Be2 Ba5 13. O-O Nxe5 14. Nxe5 ({ Ortueta could have had a good game with} 14. Ng5 {either winning an exchange or enabling the (computer) manoever} Rh4 15. Qe1 Rh6 16. Qg3 Nbc6 17. Bf4 Qf6 18. Nd6 {and White has excellent compensation for the two lost pawns.}) 14... Rxe5 15. Bf4 Rf5 16. Bd3 Rf6 17. Qc2 h6 18. Be5 Nd7 $1 ({Sanz finds the only defense, as for example} 18... Rxf1+ 19. Rxf1 Qe7 20. Bh7+ Kh8 21. Nd6 { loses immediately.}) 19. Bxf6 Nxf6 20. Rxf6 $1 ({Ortueta finds the best continuation. After a normal move like} 20. c4 d4 {he would fight with an exchange against two pawns. So the sacrifice to keep the initiative is the best idea in this position.}) 20... Qxf6 21. Rf1 Qe7 22. Bh7+ Kh8 23. Qg6 Bd7 24. Rf7 Qg5 25. Qxg5 hxg5 26. Rxd7 Kxh7 27. Rxb7 Bb6 28. c4 ({As a result of Sanz' ingenious defense, he is a symbolic pawn up. Exchanging the light pieces with} 28. Nc7 {should lead to a draw now, but the text move also isn't wrong.}) 28... dxc4 29. Nc3 $2 ({For the first time in the exciting game, there is a real mistake made. The normal contination} 29. Nd6 {should lead to an equal game after} c3 (29... Rd8 30. Nxc4 Rd1+ 31. Kf2 Rc1 32. Nxb6 Rc2+ 33. Kf3 Rxb2 34. Rxa7 Rxb6 $11) 30. bxc3 c4+ 31. Kf1 g4 32. Nxc4 ({or} 32. Ke2 Rf8 33. Nxc4 $11) 32... Rf8+ 33. Ke2 Rf2+ 34. Kd3 Rxg2 35. Nxb6 axb6 36. Rxb6 Rxa2 37. Rxe6 Rxh2 38. Ke3 $11 {. But then, it is hardly possible to criticize Ortueta for not seeing the crazy upcoming combination.}) 29... Rd8 30. h3 Rd2 31. Na4 {[#]} Rxb2 $3 {Black gives his rook, only to be unable to decide if he should capture the white rook later and ultimately promote his pawn instead.} 32. Nxb2 c3 33. Rxb6 $1 c4 $3 (33... c2 $2 34. Nd3 axb6 {leaves only White winning chances while the text move protects d3. So for now he refuses to take the rook. However, White has another defense.}) 34. Rb4 $1 {and the rook easily stops the pawn. White wins...} a5 $3 {...but only if we ignore the a-pawn which after refusing to take the rook now decided otherwise. The fatal attack destroys the coordination of the both pieces. It is incredible how knight and rook can't stop a pawn that is only on the third (sixth) rank.} 35. Nxc4 c2 { And now the a-pawn decides otherwise again, giving his counterpart on the c-file time to promote. It is of note that, on the contrary to stylized versions of the final combinations, there is no fortress for White here, as he will lose a piece after the promotion while Black can keep at least either the a- or e-pawn.} 0-1

In the second case, I was unable to trace a full game, but it is most likely not fabricated.

[Event "Petrograd"] [Site "?"] [Date "1907.??.??"] [Round "?"] [White "Alekhine, Alexander"] [Black "Hofmeister"] [Result "1-0"] [Annotator "Hornecker,Siegfried"] [SetUp "1"] [FEN "k1rb4/p3r3/Pp1Q1p2/3P2q1/P1P4p/1R4nP/2R3PK/6B1 w - - 0 1"] [PlyCount "11"] [EventDate "1907.??.??"] {White to move is a piece down. The world champion thought out a devilish combination in which Black falls, instead of drawing with the best defense.} 1. c5 b5 $2 {Instead of this weak move, pretty much every sane defense would have drawn. Only one example, part of an analysis by Taylor Kingston in an in-depth analysis of Alekhine's book of his best games, is given.} (1... Re2 $1 2. Rxe2 Nxe2 3. Qd7 $1 Qe5+ 4. Kh1 Qc7 5. Qxc7 $1 Bxc7 6. d6 Bd8 7. cxb6 Bxb6 ({ very cool is} 7... Rc1 8. b7+ Kb8 9. Kh2 Nxg1 10. Re3 Rc6 11. Re8 Nf3+ 12. gxf3 Rxd6 $11 {with a position that resembles a study.}) 8. Bxb6 axb6 9. Rb4 $1 (9. Rxb6 $2 Rc1+ 10. Kh2 Ng3 {and Rh1 mate to follow.}) 9... Rc6 10. Rxh4 Rxd6 { "and no win for either side is apparent." (Taylor Kingston)}) 2. axb5 Ne4 ({ Extremely beautiful and study-like is the variation} 2... Nf1+ 3. Kh1 Ng3+ 4. Rxg3 Qxg3 5. b6 Qxd6 6. cxd6 Rxc2 7. dxe7 Bxe7 8. b7+ Kb8 9. Bh2+ Rc7 10. Bf4 $3 {[#]and White wins, for example} Bd8 11. g3 f5 12. Kg2 Be7 13. gxh4 Bxh4 14. Kf3 Be7 (14... Be1 15. Bxc7+ Kxc7 16. d6+ Kb8 17. Kf4 $18) 15. Ke3 Bc5+ 16. Ke2 Be7 17. d6 Rc2+ 18. Kd3 Bxd6 19. Bxd6+ $1 Rc7 20. Ke3 f4+ 21. Ke4 f3 22. Bxc7+ Kxc7 23. Kxf3 {and the rest is agony.}) 3. b6 $3 Nxd6 $2 ({A more ingenious defense was} 3... Bc7 {when White needs to find} 4. b7+ Kb8 5. bxc8=Q+ Kxc8 6. Kh1 $3 Bxd6 7. cxd6+ Kd7 8. dxe7 Ng3+ 9. Rxg3 Qxg3 10. Rb2 Kxe7 11. Rb7+ Kf8 ( 11... Kd6 12. Bh2 $18) 12. Rxa7 {and the a-pawn decides. This would have been hard to evaluate in a game.}) 4. cxd6 Rec7 ({Black plays an aesthetically pleasing move, whereas} 4... Rxc2 5. b7+ Kb8 6. Bxa7+ Kxa7 7. b8=Q+ Kxa6 8. Qb5+ Ka7 9. Qa4+ Ba5 10. Qxa5# {would have just be a normal finish. What pawn should take on c7 in the main variation now? The answer is astonishing...}) 5. b7+ $1 Kb8 6. d7 $3 {[#]None of the pawns takes on c7. I only saw this combination recently in a book, and who knows how many other such incredible combinations I might not know about that already have happened in a game... Black, of course, gave up here.} 1-0

But is there anything a pawn cannot do? Well, there is a rule that forbids pawns from moving backwards, but what if I would tell you that even this restriction does not need to apply always in the magical world of chess?

This will be a controversial entry, at least for the study experts. In a recent issue of the British magazine "eg" I have proved this study to be unsound after the critical point, owing to a dual it has. But as this is aimed at players mostly, at people who will enjoy an incredible combination even if there was another way to win, and as the first 17 moves are correct, I believe that the idea behind this study makes it worthy to be seen. You have seen a pawn promote to a queen, possibly even to a knight – in fact, some very famous games (as two examples you can look up I will only name Suat Atalik vs. Tony Miles, 1993; Runau vs Schmidt, Germany 1972) end with that way. You will be aware that a pawn can promote to a bishop or rook as well. But what if a pawn's promotion is only a disguise to hide its inability to make an otherwise impossible move? What if a pawn needs to make just one step to win, but that step... is backwards? s

The curious position sees White in constant danger of Black's "Siegfried rook" ideas, which is the term for a piece that sacrifices itself in every way possible, stemming from the German Ring Cycle (Der Ring der Nibelungen, to be precise), not to be confused with Tolkien's other Ring saga. The only way to break through Black's defense is by threatening a checkmate, which is possible only with a knight on b7 or c6, so the first move promotes accordingly.

[Event "Shakhmaty v SSSR, 2nd Prize"] [Site "?"] [Date "1975.??.??"] [Round "?"] [White "Gurgenidze, David"] [Black "Kalandadze, Velimir"] [Result "1-0"] [Annotator "Hornecker,Siegfried"] [SetUp "1"] [FEN "8/p3P3/rp6/kp6/1p4PK/1P6/1P5p/4r2R w - - 0 1"] [PlyCount "41"] [EventDate "1975.??.??"] 1. e8=N (1. e8=Q $2 Rxh1 {on the contrary leaves White with a queen that can't do anything.}) {[%tqu "Please take a careful look at the diagram and try to find a winning idea for White. How can he hide the king from eventual checks? The winning idea is to hide on g2(!), but this requires White to play 2.g4-g3 Rf1 3.Nd6 Rf4+ 4.Kg2, winning. There is only one issue: A pawn never can move backwards... so White for now will go for a pawn promotion as everything else won't make any progress. We'll figure out later how to win.","","",Rg1,"",10, Rxh1,"",0]} 1... Rg1 $1 ({Black has to go for the "Siegfried" defense, as now} 1... Rxh1 2. Nd6 {leads to an early checkmate on b7.}) {If White could move his pawn from g4 to g3 it would provide a shield for the white king, and that would force mate against the black king. But pawns can only move forward. White has to promote the pawn – carefully – and then move it to the square where he would have liked the pawn to be: g3.} 2. Kh5 $1 Rf1 ({The same result is gotten if the rook goes to any other square on the first rank between c1 and f1. Of course he will always be unable to leave the first rank, as for example} 2... Re1 3. g5 Rxe8 4. Ra1# {demonstrates.}) 3. g5 $1 Rg1 {With the repeating manoeuver White will be able to advance his king and pawn further up the g- and h-file, all the while Black is bound to both the first rank and to being able to leave it, as capturing the White rook at any point leads to the mate on b7. On the other hand, White also has no time to spare, as the Black rook tries to sacrifice itself. So this manoeuver repeats a few more times.} 4. Kh6 Rf1 5. g6 Rg1 6. Kh7 Rf1 7. g7 Rg1 8. Kh8 Rf1 {The time has come for White to promote the pawn. What should he choose? A knight or bishop won't be able to shield the king, so only a rook or queen will do.} 9. g8=Q ({or} 9. g8=R) 9... Rg1 {And now the queen must shield the king, but after any move the rook will go to f1, hindering progress. The only sensible move is...} 10. Qg7 $1 ( 10. Qg6 $2 Rf1) 10... Rf1 11. Kh7 Rg1 {Of course, this manoeuver also can be repeated a few times.} 12. Qg6 ({Shielding the king with the knight is too clever:} 12. Qg2 Rf1 13. Ng7 Rxh1 14. Qxh1 {is only a stalemate, while the knight is too far away to deliver any checkmate in two.}) 12... Rf1 13. Kh6 Rg1 14. Qg5 Rf1 15. Kh5 Rg1 16. Qg4 Rf1 17. Kh4 Rg1 {[#] Do you remember that I said a pawn can't move backwards? I lied - but only in this very special case!} 18. Qg3 $1 {White just has played 2.g4-g3. All it took to make an impossible move legal was to promote the pawn to a queen. Chess can be so simple and yet complicated sometimes. An Indian proverb calls chess "a sea in which a fly can bathe and an elephant drown". Now the full extent of the promotion is clear: As Tim Krabbé explained, White didn't promote to a queen but rather to a non-existant piece - a pawn that moves backwards!} ({As I showed in "eg" 201 in July 2015, White can lose a tempo here:} 18. Qg2 Rd1 19. Qg3 Rd4+ {. But this does not make the study any worse.}) 18... Rf1 19. Nd6 $1 ({White can also win with} 19. Nc7 $1 Rf4+ 20. Kh3 Rf1 (20... Rh4+ 21. Kg2) 21. Nd5 Rg1 22. Kh4 Rf1 23. Qg4 Rg1 24. Qxb4# {. This is marked as a serious dual in Harold van der Heijden's database, but I tend to disagree. The main point of the study is correct, even if it has a possibility to lose a move at the end, and moving the knight to c7 is just another way to win after that point is shown.}) 19... Rf4+ 20. Kh3 Rh4+ 21. Kg2 {and Black is checkmated the next move.} 1-0

David Gurgenidze (b. 1953) and Velimir Kalandadze (b. 1935) are two prolific chess composers from Georgia. Together with the late Iuri Akobia (1937-2014) and Gia Nadareishvili (1921-1991), probably among others, they shaped study composition in Georgia, making malyutki (studies with five or less pieces) and studies with the material of kings, rooks and pawns two staples of the country. Akobia was a radio communications engineer, Gurgenidze works in the Georgian Chess Federation, Kalandadze is a former engineer. Nadareishvili was a renown psychiatrist in a hospital in Tbilisi. Nadareishvili (1980) and Gurgenidze (1990) received the title of “Grandmaster for Chess Composition”.

Before we close this episode, we would like to take the opportunity to invite readers to ask Siegfried Hornecker questions about chess composition in an upcoming IAMA on Reddit, on 12th March 2017, 19:00h European time (1 p.m. EST).

About the author

Siegfried Hornecker (*1986) is a German chess composer and member of the World Federation for Chess Composition, subcommitee for endgame studies. His autobiographical book "Weltenfern" (in English only) can be found on the ARVES website. He will present an interesting endgame study with detailed explanation each month.

Siegfried (*1986) is a German chess composer and member of the World Federation for Chess Composition, subcommitee for endgame studies. His autobiographical book "Weltenfern" (in English only) can be found on the ARVES website. He presents an interesting endgame study with detailed explanation each month.


Rules for reader comments


Not registered yet? Register