Home Page arrow Chess Life Online arrow 2008 arrow September arrow Compose Like Mozart
Compose Like Mozart Print E-mail
By Gary Kevin Ware   
September 25, 2008
 In my previous article, about how to solve chess problems, Start Solving Now Part I and Part II, aside from using my own experience, I distilled the wisdom of some of the great composers, to help us learn how to solve chess problems better. In this case, I admit that I have no personal, practical experience in composing problems, but again, I will distill the wisdom of some of the great composers who have written about how to compose chess problems, such as Sam Loyd, Pal Benko, Brian Harley, and others.

Before one can compose a chess problem, first one must know what a chess problem is. For instance, how does it differ from a game position or even an endgame study? In an endgame study, the solver has to demonstrate a win or draw for White against a superior or equal force and is allowed an indefinite number of moves to do so. In a composed problem, a mate must be shown against any defense in a stipulated number of moves. In an endgame, the solver is fighting against material odds; in a problem, he is fighting against time.
In, Mate in Two Moves: The Two Move Chess Problem Made Easy by Brian Harley, he defines a chess problem as "a position constructed to display, to best advantage, an idea (or combination of ideas) that lead to a forced mate in a definite number of moves." Although the composition of chess problems dates back over a thousand years into the Middle Ages, the foundations of the problem of today were laid over a century ago by the composers of what is now called The Old School. The problems of these composers were to a large extent based on forced sacrifices and checking continuations running into many moves, with little or no choice of defensive play. Gamelike positions were regarded as so desirable a feature that pieces, which took no part in the play, were frequently added to give the appearance of relatively equal forces.

This period was followed by The Transition School, on which H.G.M. Weenink (1892-1931) comments in The Chess Problem: "The period between 1845 and 1862 is of the utmost importance in the history of the development of the chess problem." When in 1845, Howard Staunton published Reverend Loveday's Indian Theme problem (see C1, 64 Square Problem Tour ) featuring critical play, this stimulated the imagination of problemists, who soon invented other kinds of line-themes such as cutting-point themes, doubling themes, Bristol Clearance, decoy themes, focal play and introduced tasks such as Albino and Excelsior.

Once a definite study of specific themes was undertaken, it gradually came to be recognized that aesthetically it was desirable to illustrate them without unnecessary moves. It is inartistic to take six or seven moves to show a theme, which requires only three for its presentation. Now, problems are nearly all mates in two or three moves; mates in four moves or longer are now only composed where the character of the theme requires that length for its development.

The precepts of The Old German School, also called The Continental School, were advocated for half a century in the German periodical, Deutsches Schachzeitung, founded in 1846. Dr. Johann N. Berger (1845-1933), who became the 'high priest' of the school, formulated its ideals in a book published in 1884, entitled Das Schachproblem, und dessen kunstgerechte Darstellung. He was so dogmatic that he enunciated his views as the rigid laws of a code, which the aforementioned Weenink synopsized as follows:

1. Pure mates are essential in the principal variations.
2. The mates shall be economical, the pieces developing their full activities (for instance, the queen shall not play and mate as a bishop).
3. White's continuations in the principal lines shall be quiet.
4. No more moves shall be introduced than are required for the expression of the theme.
5.There shall be no short threats; no checking threats; if need be a threat with quiet moves throughout, but by preference waiting moves.
6. Variations and tries must be introduced; problems containing duals should be reconstructed to embody the dual play as additional variations.
7. White's moves should increase in strength with the progress of the solution, captures and flight-taking moves being avoided in the keys.
8. Problem-schemes born in embryo in the imagination of the composer must be developed and completed according to the problem Laws.
9. Problems should present a game-like position, ample freedom of attack, and a certain measure of difficulty.

As Weenink comments, the only part of these "Commandments" which has passed "entirely into the discard is the first clause of No. 9. The others are more or less important still, and are still aimed at by great numbers of composers."

In Pal Benko: My Life, Games and Compositions, Benko gives his basic rules of problems which are more concise and less dogmatic: "A problem should express one or more ideas known as "themes". It must be original or at least an improvement on a known theme (or a combination of several themes). It must be sound- only the stated solution should work. The following rules apply to the technical aspects: Economy: Every piece must serve a purpose- the fewer pieces you use, the better the composition. Variations: The more "theme-related" variations you offer, the better the composition. Difficulty: The solver should be forced to work hard for the solution, though this is not a top priority."

As to my title, in the same book, Benko offers the following anecdote, "A young musician who longed to create beautiful music once asked Mozart how to become a great composer. "It is first necessary to play an instrument well," said Mozart. "Of course, one must also learn the basic rules of composing and be familiar with the work of all the great composers of the past." "But you were only six when you began to write music!" interjected the young musician. Mozart replied: "That's true, but I never thought to ask anyone as you're doing now."

The first example  that we'll look at comes from Thomas Taverner's Chess Problems Made Easy: How to Solve- How to Compose.

"Let us now attempt the construction of a simple theme Two-move Problem with a R sacrifice, the concession of a flight square, and, as nearly as may be, complete economy."


Taverner explains the construction of the key, Re4 in an unusually in-depth way. His remarks are of such interest to this article, that I have quoted him at great length below (his quotes are in italics) and in several other problems. If you haven't had enough, I encourage you to try to check out the entire book, which is available for viewing on chessmaniac.com or for download (pdf) here.

The key is to be Re4...It will be noted that the other squares have been so covered that, when KxR, White will be able to mate by Bc6. We note that the P at b4 alone fails to share in the mate. We then see that if the P at f4 is moved to d4, we can dispense with the one at b4, save a piece, and bring about a perfectly pure and economical mate. But this faces us with the fact that, after our Key move, the Black K, refusing our sacrifice, may now move to c4. If b4 is protected, the N, relieved for the moment of guarding d6, and having the new P at d4 protected by the R, may move to e3 delivering mate. A White P at a3 would suffice; but we shall never compose good problems if we are content to take the easiest line.

It is desirable wherever possible to make Black contribute to his own defeat, like a Black P at b4. But it threatens to check and, as the Key is to be a waiter, its move would have to be accounted for. If the pawns were moving sideways in relation to our present position the new Black P would on its movement block a square and allow a fresh mate by Re5. In order to bring this about, give the board a quarter turn. It will often be found that this expedient will afford the way out of difficulty and lead to improvement. There are quite as many cases where the same result is brought about by giving a half turn and allowing the pawns to move in a direction opposite to that on which they at first set out. When we now place a Black P at what becomes d7, we discover that we have to add a White P at what is now f3 and remove the White P previously at g4 to f2. As f5 is now doubly guarded, we move the B to h4. The position now stands thus:


 We are assuming that the student is actually moving piece by piece as indicated and carefully noting the effects of each change. The process will give him deeper insight into composing and solving than many hours reading.

Now we must test the soundness of the position. Pf4 threatens it by checking and driving the Black K to d6, but the White R is not guarding the P, hence the N cannot mate. But Re4+ cooks the position, for, on K moving, B mates at g3 or e7. If we place the White B at d8 and the White K at e2, removing the White pawns from c2 and f2, and adding a White P at b4, we not only avert the second solution but improve the problem. It is now, the R being transferred to h4, as follows:


We now note that the Black P, besides being essential to the solution, and leading to a variation, (Pd6, Re4 mate), prevents a cook by Rh5, for after Kf4, Pd6 defeats Bc7. It was the possibility of this threat, which decided the final position of the R. It could not make the threat if it was at c4 and there would be a cook if it was at g4 (by Rg5+)."

Unfortunately, the problem is still cooked by 1 Be7 d6 2 Re4#. But by moving the R to d1, we can still retain the key of
1 Rd4 Kxd4 2 Bf6#
      Kd6  2 Nc4#
      d6   2 Re4#
Thomas Taverner (corrected by Gary Kevin Ware 2008)

Next I'd like to quote Taverner explaining another problem, this time on a two-mover based on what we call the "half-pin," a pin in which two defending pieces are alternately held by a pin, on either moving.


The idea is that, on either N moving, the R shall be enabled to mate at h4 or c7. The only moves to prevent this, after provision had been made effectually to cover all the squares would be Nd4 or Nc5. By commanding these squares by White Ns, say at c2 and b7, each of these refractory moves of the N would be met (Nd4, Ne3; Nc5, Nd6). It will then be perceived that when the N at d5 moves, it leaves a square vacant, which we see no way of covering except by replacing the B by the White Q and placing the freed B at g2.


The mates have been brought about as intended, but there are bad duals. If either Nc7, RxN or Rh4#. If Nd4, then Ne3 or Na3#. The idleness of the B is also objectionable. A Black P at d3 instead of the White one at e2 would do very well but for PxN. Here again, the expedient of a quarter turn of the board helps, because B mates after move of the Black P. But the duals with the R must be cut out, and it would be infinitely better if the Q could be behind the R so that its pin would only be unmasked just when wanted. Here, the further resource of bodily moving the position, this time two squares upward and one to the left, may be exploited, the White K being taken out of the scheme, the Ns, the B and the R being moved relatively, and a White B having to be used as cover for what is now Kb8, as a pawn would now produce a triple after, say, Nd4 by Rg8 or Pa8 (a R or Q).

The great advantage of the new arrangement is that the key may now be the at first apparently purposeless one of Q behind the R. It also prevents a dual after Nc7 as N remaining in position is not pinned and White can only mate by Ne7. It being impossible to utilize the White K actively, it is placed where it will add to positional neatness, which is always to be aimed at.


We make no claim for the position, except that it illustrates the half-pin. It will, however, stand as a lesson in composition and, in that it suggests how a piece moved behind three others may effectively attack the opposing King, also be helpful to solvers, who may take it that wherever three pieces are on the same diagonal, one being White and free to move, it is worthwhile playing behind the free piece, any piece which will command that diagonal.

In our final excerpts from Taverner, we proceed to his exegesis of two three-movers.  

The idea to be expressed is that of alternate pure mates on different squares by Knights, with a mate by one of the other pieces which may be necessary in carrying out the conception.


White, having made his initial move and Black having replied by moving some free piece, to be added later, the threat is Nc6+; K moves, Nb6#. But instead of moving a free piece, the Black K might go to e5. A Black P at f4 would enable us to meet this if we placed our free piece, say a Black P, at g7. We could then follow Ke4 by Qd5+. Then, on Kf6, Qf5#. But we note that if K played to c4, there would be a triple by N to either b6+ or d6+ or Nc6. This might be remedied by moving the White K to, say, b7; but, then, after Kc4, KxP would lead to failure. Let us try the experiment of placing a Black P at c5, removing the White P from b4. Another Black P would then be necassary at a4, because otherwise, after Kc4 and White's reply Nd+, the Black K escapes at a4.


Most of what we set out to accomplish has been brought about; but we find that c4 defeats our threat. We then note that, after this move, we should have Nc6, then (following Kc5) Qh5 would make a nice and unexpected mate, were it not that the P at g7 could interpose. A White P at h6 and the removal of the Black P at g7 (leaving the P at a4 to be the free piece) would suffice; but, noticing then that White's b6 would be doubly guarded, we place the new White P at b5 and the White K at h8. Now, after c4, Nc6+; K moves, Qh5, giving us an attractive pure mate, which certainly adds point to the position. Looking for a suitable Key, we place the N now on e7 at g6 so that it has the merit of opening a flight square on a diagonal as well as leaving that already existing on the rank on which the Black K stands.


In solving, the first thing to be noted is that the Black K may move to c4. Making that move, Nd6+ promises something, but there is no prospect after Kb4. Nb6 might be tried as an opening because it leaves the possibility of a check by the Q. As a rule, however, such a move may be disregarded as in bad form- a flight square being taken. The trial move suggests Ne5. A look around following the capture of the N shows that if the N were on a square, which would command d5, mate would be forced.  This will lead the observant to the Key. The added Black P at c5 is the most attractive and puzzling feature, a fact which should be carefully noted for future service.

1 Nge7 c4 2 Nc6+ Kc5 3 Qh5#
       a3 2 Nc6+ Kc4 3 Nd6#
       Ke5 2 Qd5+ Kf6 3 Qf5#
       Kc4 2 Nd6+ any 3 Nc6#

With a view to giving the student an insight into the composition of a more elaborate Three-mover, we next proceed to the expression of an idea involving the sacrifice of a R and a Q.


The key is to be Rc5, while the threat, on a free piece (to be decided later) being introduced, is RxR+, KxR, followed by Qd6#. If after the Key move is made, the Black N moves to c3, the threat is thwarted, but BxN+, KxR, followed by Qc7#. If RxR, we purpose playing Qe3+, so that on Black replying KxQ, BxR#. We find that to effect this we must add a White P at g3. We note, too, that the Black K may move to either d5 or e5. Let us try Ps at f5 and g5, then QxR mates after the moves referred to.


After the key move is made, Re5 and Rxf5 are open to Black. RxR would meet the first, but Black moves Nd2. It would be possible to bring off a R sacrifice and mate after RxP, if c3 were filled by one of Black's pawns and if this were done, the Black N could be done away with, because then the White B could only effectively check at c3 after the Black P had moved. So we remove the Black N, place the Black P now at b5 at c4 and remove the White P from b3 to a4. Now, after RxP, RxP+; KxR, QxP#, and if we place a White P at a2, we mate prettily.
But now, consequent on the removal of the Black N, we are confronted with the fact that Re5 or RxP are met by the short mate, Qd2. So place a Black P at d3, removing the White P at e2, then substitute the White K with a White B at d1, enabling us to do away with the P at a2 and also cover f3, and utilize the White K so as to remove the P at g5. A Black P is also required at d2 as that square must be covered for the mate on the Q sacrifice and another at f7 to avert the threatened check from the R and to act as free piece for the threat variation (RxR). Placing the White R back at c6, and testing for cooks, we find that a Black P is necessary at g4, to prevent Bf3, which would lead to mate in three. There is a try by Rd6, which introduces the intended threat and, if RxR, Q replies by taking the R and mating at c5. It was so difficult to remedy this, that we breathed a sigh of relief on striking the defense, c3, which makes this unintended threat innocuous.


From a solving point of view, the freedom of the Black R will at once suggest that an initial maneuver which only provides for a quiet move ( a second move which does not check or make an important capture) will not suffice for the Key. That the White R is likeliest will be apparent after thinking over the moves of the Q. The fact that the R cannot be attacked without the assailing piece being captured, may puzzle; but solvers must learn to suspect such possibilities as parts of the scheme. Once he discovers the threat, he must hold on, and not be disconcerted by RxR. The possibility of sacrificing the Q effectively must always be borne in mind. In this case, the point that if the Q can be so maneuvered as to command the Black R after it has taken the White one should show the way. The rest is a matter of careful analysis in which the solver should not allow himself to be sidetracked by difficulties, at any rate, until he has proved that they cannot be overcome.

1 Rc5 Rxc5 2 Qe3+
      Re5  2 Rxe5 any 3 Qxe4#
      Rxf5 2 Rxc4+ Kxc4 3 Qxe4#
      c3   2 Bxc3+ Kxc5 3 Qc7#

Now we will look at an example of problem construction by Sam Loyd, as given in Alain C. White's Sam Loyd and His Chess Problems. After Loyd had mastered Loveday's Indian (see C1, 64 Square Problem Tour ) and analyzed its theme, he lost no time in beginning experiments himself. Variations from the Loveday motive began to suggest themselves. The problem that we will consider omits the customary discovered mate of the Indian "and yet furnishes a new and interesting branch of the family." This position has proved one of Loyd's most suggestive works, and from it has sprung a whole group of problems of the Passive Sacrifice type. By a passive sacrifice is meant the surrender of its powers and activities by any piece (usually White), without its actual removal from the board by capture. There are several kinds of passive sacrifice, the intersectional (or Indian) type being the most ingenious and also the most popular.

  "To give a faint conception of the vagaries that flit through the mind of a composer- My crude idea for rearing a little Indian was this:


(#3, 1 Bb1; 2 Nc2) A slight examination revealed several mates in one, many in two, and more in three. To prevent the mates with the knight and to take away the dangerous guard of f7, I substituted a Pawn for the Knight:


(#3, 1 Bb1; 2 Nc2) To cure the radical weakness of 1 Kh6, I turned the problem on its head:


(#3, 1 Bh5; 2 Ng4) The forces were still so strong that I could not prevent other solutions, although I tried a thousand expedients before essaying this new departure:


(#3, 1 Ba8; 2 Nc6) I tried many ways to prevent the mate resulting from 1 Ke2, but finally had to employ a Knight instead of a Bishop, and also raised the pieces to prevent the other key of 1 Bb7:

(#3, 1 Bb8; 2 Nc7) This was better, but I was compelled to make another change to prevent 1 Qxg5+:


(#3, 1 Ba8; 2 Nb7) There were still other solutions, 1 Kf2 and 1 Qg6, so I added Black pawns at e3 and h7, and a White Pawn at h6. But I found no way of correcting the move 1 Nd4, except by dispensing with the Knight altogether:


(#3, 1 Ba8; 2 Nb7) Here again there was a pretty flaw, 1 Be4; 2 Nf7, so I was compelled to lower everything in order to add an extra Black Pawn, as shown in our final diagram, which makes the position sound."

Sam Loyd "Happy Thoughts" Cleveland Leader Tourney August 24th, 1876

1 Ba7 f4 2 Nb6 Ke3 3 Qd3#
      Ke4 2 Qg3

 I hope that those of you had no interest or experience in composing problems will now try your hand at it and that those who did have some interest and experience also learned something. As is always the case, I, in doing the research for these articles, have also learned something. I am considering doing a future article featuring the compositions of amateur composers, who have hopefully gained something from this article. Please submit your problems to [email protected].