Home Learn to Play Chess Improve Your Game Chess History Chess for Fun Chess Blog
|Relative Value of Pieces and Principles of Play|
|From The Modern Chess Instructor by Wilhelm Steinitz, 1st World Chess Champion.|
Wilhelm Steinitz first published The Modern Chess Instructor in 1889. The book's seven chapters were:
On this page, we reproduce Steinitz's Chapter 7, 'Relative Value of Pieces and Principles of Play'. For the About Chess introduction to the 'Relative Value of Chess Pieces', click here.
Relative Value of Pieces and Principles of Play
One of the most important exigencies in the conduct of the game is the exercise of the most critical judgment in estimating the relative value of the pawns and pieces which must be strongly taken into consideration if effecting exchanges, as well as in the formulation of general principles for the guidance of play in all parts of the game. But owing to the endless number and variety of combinations that are possible over the board, it has been found impossible to give more than an approximate theoretical and practical comparison of the relative powers of the men.
In Staunton's Handbook, page 34, it is stated that some scientists have calculated the approximate mathematical value, to be as follows: Taking the pawn as the unit, the Knight is worth 3.05; the Bishop 3.50; the Rook 5.48; and the Queen 9.94. On this basis, which in the main is in accordance with our own experience and observations, we shall proceed to indicate, in connection with the above approximate valuation, some of the most important general principles of regulating the actions of the men which we believe are now mostly accepted by the strongest masters of the day, and the knowledge of which very often enables the player to dispense with analysis, or at any rate greatly assists his calculations.
As, however, already explained in our preface, the scope of this work will not enable us to illustrate the application of our guiding maxims any further than is done in our notes to our analysis and selected games. We shall now endeavor to describe seriatim and briefly the power of each man and its most favorable mode of development, as well as to offer some hints as far as practicable about its value and action in the middle game and in the ending.
The King is considered invaluable, according to all authorities, on account of his not being liable to capture or exchange, which also involves the complication of his having to move out of check, or to cover the same, or to capture a checking man to the exclusion of the choice of other moves. Baron von Heydebrand in Bilguer's Handbuch very properly describes the power of the King for the pawn ending as stronger than any minor piece, namely, Knight or a Bishop. We are inclined to extend this valuation to all parts of the game, and we would add that the action of the King combined with one defended pawn is about equal to that of a Rook, provided that neither the adverse King nor any other hostile man can cooperate with the latter.
We agree in the main with the authorities who recommend that the King should as a rule castle early on the Kingside, but this refuge of the King is sometimes fraught with danger when one of the pawns on the King's wing -- more especially the g-pawn or h-pawn -- have been previously moved or may soon be compelled to advance. Likewise when the opponent has obtained the majority of pawns on the Queenside it is generally better not to widen the distance between the King and the adverse majority, as the King is a powerful piece in the ending for stopping the hostile pawns.
In either of these cases it is desirable to aim first at an exchange of Queens and some minor pieces and to postpone castling or not to castle at all. The King is sometimes brought into play at d2 after developing the minor pieces on the Queenside, or at f2 after the advance of the f-pawn.
Castling on the Queenside is not often advantageous, for it leaves the a-pawn undefended as the Handbuch rightly points out. The notable exceptions are when the d-file has been opened for the player who castles on the Queenside, while the adversary cannot open that file; or when the pawns on the Kingside can be advanced for a strong attack with the cooperation of other pieces against the adverse King who has castled on the other side.
In castling on either side, it should be remembered, that the Rook's, Knight's, and Bishop's pawns on that wing in conjunction with a minor piece, generally a Bishop or a Knight at Bishop's third or Bishop's square (after removing respectively the King's Rook to e1 or King to f1) form an excellent protection against the larger majority of attacks that can be planned by the opponent. The advance of either of these pawns should therefore be postponed as long as possible, or else it will form an easier mark for the attack of the hostile men, and one of the minor pieces should be kept within convenient reach of Bishop's third or Bishop's square on the side on which the King has castled.
Excepting some openings that will be specially treated in this work it is rarely good play to move the King in the early part of the game. But this may be resorted to even with advantage in some cases when the opponent allows his e-pawn to be taken with a Knight in order to gain the f-pawn for it. For instance, after the moves 1.e4 e5 2.Bc4 Nf6 3.Nc3 Black though the second player may now safely reply 3...Nxe4 and allow his King to be disconcerted for a little while by the answer 4.Bxf7+, for after 4...Kxf7 5.Nxe4 d5 6.Qf3+ Kg8 7.Ng5 Qd7, the attack will be soon transferred to Black who has gained the advantage of the strong combination of two Bishops and the formation of an excellent center. Some other analogous cases arise sometimes in the opening and may be treated in a similar manner.
Occasionally it becomes necessary in the middle game either for purposes of attack or defense to remove the King from one side to the other, and sometimes by way of squares in the middle of the board. Such a movement ought only to be adopted with the greatest precautions for it generally involves the loss of costly material especially when the Queens are not yet exchanged.
But on the other hand, the strong defensive powers of the King ought to be fearlessly estimated, and when no such loss is threatened or the opponent cannot bring sufficient pieces up for the attack, it should be remembered that it requires a combination of great powers to mate the King. For instance, when he stands on any of the border squares and is not blocked by any of his own men, he can only be mated by forces that are rarely available for such a purpose in the middle game. A single piece will often cover his retreat or at least delay mating operations even against Queen and Rook combined. And when the King travels in the middle of the board without being obstructed in his moves by his own forces, it requires at least the combined strength of Queen, Rook or a minor piece, and one pawn, which, moreover, must be in a special favorable position for the purpose, in order to effect mate.
Staunton justly warns against giving useless checks, but recommends as generally good play, to give a check early in the game when by so doing the adverse King can be compelled to move and thus be deprived of the right of castling. The same authority also says: "Do not in all cases take an enemy's pawn that stands before your King – it may sometimes serve as a protection to him".
In the ending the King is a powerful piece for assisting his own pawns or stopping the adverse pawns. In trying to stop an adverse passed pawn that cannot be supported by his own King, it must be observed that the King must stand or be able to move to any square of a quadrate that can be formed by taking as a measure of one line, the number of squares from the one inclusive on which the pawn stands up to that of the top row inclusive. Thus, for instance, if White's pawn stands at a3, the four points of the quadrate are the squares at a3, a8, f8 and f3. If Black's King stands on any square from f8 up to f3, and therefore at the greatest distance between the position of the pawn and any square of the quadrate, he will still catch the pawn even if the latter has the move.
To give another illustration, we assume that White's pawn stands on a5 and in that case Black will be able to stop the pawn if he stands or can reach any square from d8 to d5. But it should be noticed that if White's pawn stands on its original square at a2, the Black King standing on the furthest file of the quadrate, namely: on any square from g8 to g2 inclusive, must have the move in order to stop the pawn, as the latter can move two squares at starting.
Likewise in any original position of the pawn, the adverse King, if standing on any square of his 7th row without at once being able to capture the pawn, must have the first move, even if he is within the quadrate in order to stop the pawn. But unless the King stands on the file in front of the pawn, the latter can never be stopped if there are more than four squares in any straight direction between the King and pawn.
The Queen is the most powerful piece on the board, and for that reason should not be subjected to attacks from inferior hostile men by being brought out early in the game. As the Handbuch points out it is dangerous, especially in the opening, to place the Queen on the same file or diagonal as the King.
Before the game is well developed, three pieces including the Rook, or two Rooks, may be given up for the Queen with advantage, but when the adverse position is well defended and the pieces can be brought into cooperation, three pieces, including one Rook, or two Rooks are superior to the Queen. Two Knights and one Bishop are generally inferior to the Queen.
The most favorable points of development for the Queen are d2 after developing the Queen's Bishop, or c2 as well as b3 after moving c2-c3. The latter development is especially attacking in forms of openings where the King's Bishop is played to c4. It is rarely good to play Qe2 or Qf3 in the opening, but such posts may sometimes be selected without disadvantage when the adversary has already played ...c7-c6 or is otherwise prevented from bringing out his Queen's Knight to c6, where the Queen would soon be attacked by ...Nd4. In some of the closed games, the Queen may be developed at a4 after moving the c-pawn in order to post the King's Rook at d1 and the Queen's Rook at c1 after developing all the minor pieces.
In openings in which the d-pawn is advanced to d4 the attack is often formed against the adverse Kingside, by placing the Queen at d3 after having maneuvered the King's Bishop on the same diagonal at c2 or b1. Another favorable post for the Queen in attacking the Kingside is at g3, and in some cases like the counterattack in the Evans Gambit declined, or the new attack adopted by the author against the French Defense, the Queen may advantageously be brought out at g4 for an early attack.
Yet a player should always be very cautious before capturing a hostile pawn or even a piece with his Queen, as situations often arise in which the Queen can be afterward blocked out and ultimately caught for inadequate material, or at any rate her return into her own camp is thus delayed until the opponent has gained time for instituting a formidable attack.
Owing to the original position of this piece, which is blocked up by its own men, and the nature of its movements, it cannot be made much use of in the early part of the game. The King's Rook is in many respects superior to the other for opening engagements on account of the earlier facilities for castling on the Kingside.
The two minor pieces on the King's wing can be sooner developed into attacking or commanding positions in most open games, whereas on the other side, the Queen has to be brought out in addition to the two minor pieces, which in their early development does not threaten much and leave the opponent the option of many more replies. In the majority of openings commencing with 1.e4, the castling on the Kingside also offers the first opportunities for opening a file for the Rook by advancing f2-f4, and this is of the utmost importance for that piece, which can only be brought into action on open files or rows.
It should be noticed that the two combined Rooks are in the most favorable position for attack and defense when doubled on an open file. One of the most powerful attacking posts for one Rook and still more for the two combined Rooks, is on the 7th row, for usually some of the pawns of the adversary are stationed on their original squares and are thus more liable to capture. Such a situation of Rooks also often forms an irresistible attack against the adverse King, which is usually confined on the front row.
Other common ways of leading the Rooks for an attack against the Kingside, is to bring one of them by way of f3 to h3, after the advance of f2-f4 and after the exchange or dislodgment of the adverse Queen's Bishop, and then the other Rook in a similar manner to g3. Such an attack, if well supported by minor pieces or the Queen, is often most formidable, but nevertheless, its prospect of success must be well weighed, for if the attack fails, the heavier pieces remain uselessly packed together on the Kingside, and the opponent has the better chance of winning if he can in the meanwhile form an attack with his pawns in the center or on the Queen's wing.
Two cooperating Rooks are stronger than the Queen when all points are well defended, but more especially when the King is well guarded against harassing checks. But it should be remembered that the Rooks are rather clumsy pieces to handle, while the agility and long range of the Queen in all directions afford for the latter many opportunities for defense and attack, especially in conjunction with one or more minor pieces.
The Handbuch remarks that the Rooks are most fitted for supporting the advance of the passed pawns, but much less strong for stopping them, whereas Queens and Bishops are powerful pieces for checking the pawns. It is therefore advisable for the party that has strong pawns to exchange Queens and Bishops and to retain the Rooks, while the contrary policy should be adopted for the defense.
The Rook is generally slightly stronger than a Knight and two pawns; while a Bishop and two pawns are in practical play a shade stronger than the Rook. A Rook and two pawns are superior to two Knights and a little better than a Knight and Bishop, but about equal with two Bishops. Two Rooks are a little stronger than two Knights and a Bishop, but slightly inferior to two Bishops and a Knight.
In all cases, however, a great deal depends on various considerations that have also to be borne in mind when a minor piece is given up for pawns, namely, the position of pawns, and whether their majority is compact on one wing or divided, whether the King can support his pawns or whether the adversary's King is nigh enough to stop them, whether or not one or more passed pawns can be formed, and whether there are other pawns on the board that are liable to capture or are well defended. It also must not be lost sight of that the party having the pawns, and provided there are no other pawns on the board, or all others can be exchanged, has the only chance of winning, whereas the party thus fighting against the pawns can only play for a draw.
In the ending when trying to advance one or more passed pawns without the King and against the adverse Rook alone, it is generally best to place the Rook behind the pawns in order not to obstruct their advance. But when fighting against hostile pawns, it is mostly advisable to attack them in the rear or to stop the one furthest advanced in the same manner.
Two passed pawns on adjoining rows will win against the Rook with or without the move when they have both reached the sixth square of their file, provided that the adverse King is at least a distance of three clear squares from the pawn next to him, and that neither pawn can be taken by the Rook at once. In a similar manner, three adjoining passed pawns on the fifth squares of their file will win against the Rook with or without the move if the adverse King is at a distance of at least four clear squares, and provided that neither pawn can be taken at once by the Rook. But it is necessary to know that if the Rook attacks any of the pawns excepting the middle one of the three, the pawn thus attacked should be given up and one of the others should be pushed, when the remaining two will secure reaching the sixth square before the adverse King comes up. If, however, the Rook attack the middle pawn that pawn must be first advanced.
The relative value of this piece has given rise to different opinions among masters and authorities. Some have shown or expressed a distinct preference for the Knight in the ending and it has also been asserted that in conjunction with Queen and Rook, the Knight is stronger than the Bishop.
But after careful consideration of the average of positions that have attracted our attention and the few exceptions positively in favor of either piece, we have come to the conclusion that the power of the Bishop corresponds for practical purposes with its estimated superior mathematical value over the Knight in the opening, and in the middle part as well as in the ending, and in the majority of combinations with other forces. The great power of the Bishop, especially in conjunction with the other Bishop for attack in all directions, as well as for the defense has been first systematically and consistently demonstrated in practice over the board by the great German master, Louis Paulsen, who may be regarded in many respects as one of the chief pioneers of the modern school.
In the opening the King's Bishop is preferable to the other on account of his usual aggressive bearing against the hostile Kingside. His best post in the development of open games is at c4, whence he is often retreated to d3 or c2 after advancing d2-d4 and c2-c3 if the opponent has castled on the Kingside. In some openings in which the adversary is enabled to bring his Knight to e5, or in closed games, or when the opponent threatens an attack on the Kingside by bringing his pieces or pawns to bear against the g4 square, the King's Bishop is sometimes better posted at e2 in order to avoid its being exchanged for a Knight or for other defensive purposes.
The Queen's Bishop is mostly developed at d2 or e3, but in some openings can be kept at home for a long time until f2-f4 can be played with advantage, and in case the adversary capture that pawn with the e-pawn, an excellent game will often be obtained by retaking with the Bishop.
As already stated it is often useful to keep the respective Bishop within reach of the Bishop's square on the side on which the King has castled. It is usually best to keep both Bishops in communication with both wings and for that reason as well as on account of the superior value of the Bishop it is very rarely of advantage to pin an adverse Knight.
Notably should the pinning of the hostile King's Knight by Bg5 be avoided excepting when some clear advantage or compensation can be perceived. For the opponent by attacking the Bishop with ...h7-h6 will either effect an exchange more favorable to himself, or the Bishop will have to retreat with great loss of time. It is generally disadvantageous to allow the Queen's Bishop to be driven back to g3 out of communication with the other wing, especially when his Knight is posted at f3.
For defensive purposes it is generally advisable to retain the Bishop of the color on which the majority of pawns are placed or likely to be fixed, more especially when such pawns are stationed on different separated diagonals. For the attack, the Bishop should be retained of that color on which the majority of the adverse pawns are placed and an advantage will then generally be effected by endeavoring to break through with well supported pawns.
The superiority of the Bishop over the Knight is also shown by the fact that the former when placed on any square of the board will command at least 7 squares of one or more clear diagonals. In the middle of the board at e4, e5, d4, or d5, he will command 13 squares. On the other hand, the action of the Knight may be reduced to the command of no more than two squares, if he be placed into any of the four corners of the board, and the maximum of squares which he can command is eight.
The great power of the two Bishops combined has already been alluded to. They are a little superior to Bishop and Knight and considerably stronger that two Knights. With the qualifications mentioned in our description of the properties of the Rook where we have also given some comparative valuations of Bishop and Rook with pawns on either side, we would further compute that two Bishops and two pawns are considerably stronger than Rook and Knight, and that one Bishop is much better than three pawns.
But it should be pointed out that two passed pawns on the sixth row even if separated will win against the Bishop with or without the move, if neither can be taken at once, and the adverse King stands at least three clear squares distant from either pawn. On the other hand, a Rook would easily stop such two or even more separated passed pawns if they cannot be supported by their King for some time, by simply placing the Rook on his second or fist row.
Some of the old authorities maintained that this peculiar piece should not be brought out in any manner as to block one of the pawns, and therefore not at Bishop's third before having advanced the respective Bishop's pawn two squares. The King's Gambit and the Bishop's Gambit are founded on that theory. But it is now universally acknowledged among experts that after 1.e4 e5 on each side, 2.Nf3 or 2.Nc3 are excellent moves, and in most openings the defense ought also to bring out the two Knights on their respective third squares without minding the blockation of the pawn in front of them.
After castling on the Kingside it is generally a good plan to remove the King's Knight in order to advance f2-f4, and often Ne1 is the best retreat for the purpose. But we disapprove on general principles of the plan sometimes adopted of playing h2-h3 in order to retreat Nh2.
The Queen's Knight is often maneuvered from c3 via e2 to g3 for the attack, but he is also developed sometimes via d2 to f1 either before or after developing the Queen's Bishop and thence to g3 or e3 with good effect. When either Knight can reach the adverse f5 without being liable to be driven away or exchanged he will occupy a very menacing position against the adverse Kingside, which will greatly strengthen any attack in that quarter.
The Knights are well adapted for entering into a "hole" or a weak square of the adverse game (of which terms we shall give some further explanations anon) especially when supported by pawns on each side.
A Knight is only very slightly stronger in general than three pawns. Of its other relative valuations we have already spoken under the previous headings, but it is a peculiar feature of the Knight that he will be generally stronger than the Bishop in the ending when the opponent has a doubled pawn that cannot be dissolved, more especially when the one in front is of the opposite color of the Bishop and is not protected by another pawn, for then the Knight by attacking that pawn will at least keep the adverse King engaged for its protection, while his own King will be free for action. This ingenious maxim was chiefly brought into recognition by Herr Winawer.
The skillful management of the pawns which form a phalanx before the King and the other pieces, is one of the most important items in the conduct of the game. Owing to the privilege of promotion to a Queen, or any other piece chosen, which the pawns possess when reaching the eighth square the loss of one of them is in the large majority of cases fatal among first-class masters. It is, moreover, now recognized among experts that not alone the weakness of one single pawn but also that of one single square into which any hostile man can be planted with commanding effect, will cause great trouble, and often the loss of the game, and that by proper management of the pawns such points of vantage need not be opened for the opponent.
The center pawns, namely, the e-pawn and d-pawn will have to be moved in the larger majority of openings sooner or later in order to free the pieces on each side, and they are not alone the best fitted for commencing operations, but we would lay it down as a rule that they are the only ones that ought to be moved in the early part of the game for various reasons.
In the first place, as long as the three pawns on each wing remain unmoved, there is no weak square or a "hole" on the side which takes that precaution. The latter term which is now generally accepted as a technical definition, was first used by the author in The International Chess Magazine of November 1886, where the disadvantage which it is intended to describe was also first pointed out, and it is most important for the learner fully to appreciate that disadvantage. The "hole" means a square on the third or forth row in front of a pawn after the two adjoining pawns have been moved or captured.
Thus, for instance, after the opening moves 1.e4 e5 2.c4 there are already two holes in White's camp, namely, one at d3 and one at d4. These holes will be all the more dangerous as long as the adverse e-pawn remains at e5 for that pawn stops the advance of two hostile ones and by skillful play Black will retain that advantage for a long time. If White's d-pawn is moved forward to d3 that pawn will be weak and even if he succeed in exchanging that pawn for another, the squares at d3 and d4 will remain weak, and White will have to guard against the entrance of hostile men on those squares with one or more pieces, since both the pawns that previously could afford protection against such entrance are advanced.
A hole or weak square are still more troublesome when the opponent is enabled to open the file on which they are situated for his Queens and Rooks. In the opening or middle part a hole or weak square are most dangerous in the center or on the Kingside before Queens are exchanged, but in the ending such weak points are generally more troublesome on the Queenside.
In the next place, it is a great advantage for the ending to have as many pawns as possible unmoved on their original squares, for it is often most important to be able to gain a move by having the option of pushing a pawn one or two squares. Furthermore, we have already explained that three unmoved pawns on the Kingside in conjunction with a minor piece form a strong bulwark against an attack on that wing, and we shall also show anon some reasons against moving the pawns on the other wing.
Staunton's Handbook, page 44, gives the following good advice: "It is generally advantageous for your pawns to occupy the middle of the board, because when there they greatly retard the movements of the opposing forces. The e-pawn and the d-pawn at their fourth squares are well posted, but it is not easy to maintain them in that position, and if you are driven to advance one of them, the power of both is much diminished."
To this we would add that in general two pawns are stronger abreast than on a diagonal. The former command two black squares and two white ones in front, while in the latter situation, one of the squares is occupied by a pawn and all the points covered are only of one color.
As a rule it is unadvisable to advance any pawn beyond the fourth square, for the further a pawn is advanced into the hostile camp the sooner he becomes liable to capture or inconvenient attack especially in the end. At the utmost a pawn may be sometime advanced to the fifth square when he can be well supported on each side by so-called chains of pawns that cannot be broken up, but it is very rarely good play to advance a pawn to his sixth square.
In the early part of the game the formation of a center such as two pawns abreast at e4 and d4 is a very desirable object, and in the Gambits of the Kingside the f-pawn is even sacrificed for that purpose. With the view of strengthening the center it is usually better to capture with a pawn toward the middle rather than toward the wing when the capture can be effected by two different pawns.
When both sides have moved 1.e4 e5 and have also castled on the Kingside, it will be often advantageous to allow the f-pawns to be doubled in order to form some attack on the open g-file, or else with the object of afterward dissolving the doubled pawn by advancing the pawn f3-f4. In like manner, the doubling of a pawn on the c-file may be useful in order to obtain command for the Queen's Rook on the open b-file and with the view of advancing c3-c4. But an isolated doubled pawn, especially one on the Rook's file, is mostly a great disadvantage.
Most particular care should be taken that the opponent does not obtain the majority of pawns on the Queenside, on the wing opposite on which the Kings of both parties usually castle. For a skillful player will generally manage to cut off the King from crossing to the other side, and the weaker pawns, thus deprived of the help of a powerful piece, will rarely be able to offer sufficient resistance to the opposite superiority of force. The majority of pawns on the Kingside is rarely of much use, for the pawns of that wing cannot well advance without exposing their own King, and in the ending the hostile King is near at hand for stopping them.
Each pawn has its own peculiarities which we shall endeavor to describe briefly.
The two Rook's pawns are the weakest, as each only commands one square, while the others command two. But each when advanced is only liable to be attacked by one pawn on the hostile Knight's file, while the other pawns can be attacked by two hostile pawns, one on each side. When the opponent has first moved ...h7-h6 after castling on the Kingside while you have not yet castled, you may also reply h2-h3 with the view then of advancing g2-g4 and endeavoring to break through with the pawns on that wing. It is also good play to drive back a hostile piece by h2-h3, but otherwise, especially when you have castled Kingside such an advance is not good, for it exposes that pawn to attack in many contingencies and it also makes it inconvenient to advance the f-pawn, since a hole is then formed at g3.
The g-pawn if advanced to g3 leaves at once a hole at h3 and f3, for it is assumed that the e-pawn has already moved, or will have to move soon. If he advance to g4, supported by h2-h3, he leaves additional holes at f4 and h4.
It is advantageous to advance the f-pawn to f4 after castling when an adverse pawn is fixed at e5 by your own e-pawn which should be well defended. If your d-pawn has been exchanged for the opposite e-pawn, it is more often better to play f2-f3 in support of your e-pawn. If the e-pawn has been exchanged on each side, it is rarely good to advance the f-pawn, for it leaves a weak square at e3 against which an attack of the hostile Rook can also be directed. If the f-pawn remains unmoved, he will often give good support to the Queen's Bishop or King's Rook at e3.
The advance of the e-pawn to the fifth square is specially objectionable, as the opponent will mostly gain opportunities, by ...f7-f6, of opening an important file for his Rook. Likewise, if the d-pawn play to his fifth, the answer ...c7-c6 will release the adverse Queen and open a promising file for the hostile Queen's Rook.
When the d-pawn has been exchanged it is seldom right to advance c2-c3. Likewise, when the d-pawn is still at d3 the advance of the c-pawn will leave the d-pawn weak, and again, under other conditions, it retards the development of Nc3 with scarcely enough object in the opening. But still c2-c3 is often a good move latter on.
The advance of the b-pawn naturally leaves holes at once at a3 and c3, as d2-d4 or d2-d3 are either supposed to be done already or sure to follow. Finally the early pushing of a2-a3 can hardly do any good, but loses time and makes the subsequent advance of c2-c4 which is sometimes good and necessary, objectionable on the ground that a hole will be created at b3.
Thus it may be repeated in general that in most openings only the e-pawn and d-pawn should be maneuvered in conjunction with a rapid development of the minor pieces, and though the f-pawn and the c-pawn may sometimes assist, it is at least useless and often compromising to move Rook's pawn or Knight's pawn on either side in the early part of the game. A pawn attack may, however, often be formed with advantage when the opponent has crowded too many pieces on one wing or when he has given an opportunity for effecting a promising break through on either side by advancing one of his pawns; but as a rule the fight in the center in conjunction with the two Bishop's pawns will be sufficient, and at least the option of moving one or two squares ought to be reserved for the ending for the other pawns.
There are other principles based on reasonings by analogies between different positions, as well as comparisons and combinations between different principles when they come in conflict with each other, but as explained in the preface they are outside of the limits of this work, for they would require too laborious illustration. However in our introductory comments of the games between Messrs. Steinitz and Tschigorin we give some instances of the application of principles in the opening with some explanations of their influence on later stages of the game.
[Text from The Collected Works of Wilhelm Steinitz courtesy of ChessCentral.com; used with permission; reformatted for the Web.]