6.Nc3It would be quite bad to play P—B 4 and try to hold the extra pawn at the expense of development. Black would very soon occupy the King's file with his Rook and there would be no time for White to bring his King into safety, e.g. 6 P—B 4, B—Q B 4; 7 Kt—K B 3, Castles; 8 B—K 2, R—K 1, and already now there are threats of Kt—K 5 or Kt—Kt 5 followed by B—B 7 ch or Kt—B 7.
7.Bd2White would of course like to continue with B—K 3 in order to make a fight for the possession of the diagonal. He would, however, lose his chance of castling through Black's Q—K 2. This is detrimental in all such cases where the lines in the centre are open or likely to be forced open at any time.
11.Be3White has not yet completed his development, and his first care should be to bring out his K Kt. This he could have done without difficulty, thus: 11 B x B, Kt x B; 12 Kt—R 3. After the move in the text, Black not only occupies the King's file but gains a move in so doing.
13…Qd6Black's course is obvious; he must win the Q P. The forces will then be equal in material, but there will remain a flaw in White's position, namely the exposed K B P, and this tells in the ending.
14.h3Now the square at K Kt 3 is unprotected, and this is serious in view of a probable Knight's ending, where, moreover, it will sooner or later be necessary to play P—K Kt 3 in order to support the K B P. Both the K Kt P and K B P would be weak, with the King on the other wing, and be under constant threat of being captured. The game does proceed as indicated, and the simple and logical manner in which Marshall brings home his advantage in a very short time shows convincingly how fatal a shattered pawn position can be for the end-game. Instead of the move in the text, White should have played B x B followed by Kt—B 3, which would have completed his development without making another pawn move.
19.g3making a loophole for the King. In this case the move is correct, as the threat of mate ties the Black Rook to his rank. It is wrong to make a loophole, as weak players are fond of doing, as early as possible “in case,” before it is shown that there will be a need for it, or that there will be a Rook ending.
20.a3White is afraid of playing 20 Q x P on account of Kt—Kt 5, which threatens Kt x B P followed by Q—Q B 3. 21 Q x P would not be a sufficient defence because of Q—B 4 threatening mate, and on the other hand 21 Q—R 4 would conjure up a dangerous attack, beginning with P—Q Kt 4. When the players castle on different wings, there is always the danger of the opponent sacrificing pawns and opening up files for his Rooks and Q against the castled King. The game then assumes a wild character, and as matters are generally settled one way or another in the middle-game, end-game considerations, both with regard to number and position of pawns, can be disregarded. Experience has shown that the player who develops his attack first is likely to win, and that it is of little use to submit tamely to an assault of this kind without attempting a counter attack. Such games are very difficult for the beginner to understand. There is about them something violent and difficult to estimate, and years of practice are necessary in order to gain the judgment required for weighing up the possibilities of attack and counter attack, where the Kings have castled on opposite wings.
25…Kf8The Black King now pushes forward irresistibly, and attacks the weakened pawns on the King's wing. The White King cannot get any nearer, as a check by the Black Kt would win a pawn at once. The end is easy.
30…h5Now the P at Kt 3 is “backward” and therefore lost.
Edward Lasker, Chess Strategy (1921) · Public domain · source