1. At the beginning of the game the forces stand in equilibrium.
2. Correct play on both sides maintains this equilibrium and leads to a drawn game.
3. Therefore a player can win only as a consequence of an error made by the opponent. (There is no such thing as a winning move.)
4. As long as the equilibrium is maintained, an attack, however skilful, cannot succeed against correct defence. Such a defence will eventually necessitate the withdrawal and regrouping of the attacking pieces and te attacker will then inevitably suffer disadvantage.
5. Therefore a player should not attack until he already has an advantage, caused by the opponent's error, that justifies the decision to attack.
6. At the beginning of the game a player should not at once seek to attack. Instead, a player should seek to disturb the equilibrium in his favour by inducing the opponent to make an error - a preliminary before attacking.
7. When a sufficient advantage has been obtained, a player must attack or the advantage will be dissipated."

From David HOOPER, Steinitz' Theory, British Chess Magazine Vol. 104, p.370 Sept 1984.

 References:

• Steinitz' International Chess Magazine
• Steinitz' Modern Chess Instructor
• Lasker's Manual of Chess
• Euwe's The Development of Chess Style

Notes

1. The theory is the antithesis of the previously held view, that inspired play might of itself create a winning position; rather, (3) winning positions can come about only through errors by the opponent. Creativity and ingenuity still have a role, in the setting of problems for the opponent that they may fail to solve.
2. The point (4), that, if a defender played correctly, then any attack commencing from an equal position can be refuted, was first indicated by Louis Paulsen. He declared (unfashionably) that all gambits can be defended, and practised what he preached; in this way Paulsen was the progenitor of the fully formed theory of Steinitz. Defensive play has its own principles:
   The defending side must be prepared to defend and make concessions, e.g making a weakening pawn move. However, the defender should avoid making concessions until forced, and then should make only the minimum concession necessary to meet the threats.

    

3. The last point (7) is attributed by Lasker to Steinitz; DH cautions that he was unable to find such a statement in Steinitz' works. Lasker's version goes on to say:
   The right to attack belongs only to that side which has a positional advantage, and this is not only a right, but also a duty, otherwise there is the risk of losing the advantage. The attack is to be directed against the weakest spot in the opposing position.
4. DH was prompted to write this article after reading Purdy's "The Great Steinitz Hoax", Chess Player's Quarterly, 1978 [which is, according to BCM, reprinted in No.2 of Lasker and his Contemporaries], in which Purdy apparently argues that Lasker was more the author of the theory than Steinitz.
   Purdy and Hooper disagree on what constitutes the theory.
5. Lasker certainly was the principal pedagogue and advocate of (his version of) Steinitz' theory through his Manual. Lasker's presentation was coloured to some extent by the principles of his own general philosophy of Struggle (Work, Economy and Justice); Lasker also offered extensions to the Steinitz theory, viz. principles of Cooperation and of Justice (in the sphere of chess).
6. The influence of the theory is so substantial that instances need hardly be found, but a couple interested me recently:
   1. Jon Speelman's introduction to his Best Games 1970-1980 collection, where he refers to the "story" style of annotation - where was the crucial error by the loser? This of course assumes that the loser must have made at least one mistake, and Speelman quite explicitly signs up to the theory that there is Justice in chess, and that to lose you must have made a mistake - or two, really, and maybe three if you are White.
   2. An annotation in Chess Monthly where the GM reluctantly admitted that despite a series of brilliant moves by Black, White was at no disadvantage. This is the corollary of the same point; in a balanced position, even brilliant play, correctly met, will lead only to another balanced position.
7. What is the significance of Steinitz' theory for the club player? Different aspects have different importances.
   1. Defensive play may be so poor that an attack undertaken with insufficient means might well succeed in practice.
   2. So, defensive play, as modelled by Paulsen and Steinitz, must become a strength of yours.
   3. Moreover, while winning combinations may theoretically be available principally (if not only) in advantageous positions, in practice tactical opportunities and especially oversights can appear at any time, so at each move you should routinely check for combinational blows for each side.
   4. It may not be so important to decide who has the advantage as to decide what to do, and then do it well!
      "The question that matters to you in actual play is simply, 'What is my best move?', and if you can decide without being sure who has the theoretical advantage, so much the better" -- PURDY