Parts 1 and 2 of this series addressed how to formulate a middlegame plan by comparing how much space each player has along each file. (Italicized terms are defined in the "Glossary of Chess Terms," accessible from this website's "This Website" menu.)

This article broadens the discussion to include pawn islands, pawn majorities, and holes.

The game Erdai-Levin (National Chess Congress, 1992) reached the following position after 24. fxe3.

Let's consider how the above-mentioned aspects of pawn structure apply to this position. To help us focus on the pawns, here is the result of removing all of the pieces from the above position.

1. Pawn Islands - Identification

White has three pawn islands: a/b/c-pawns, e-pawn, and g/h-pawns. Black has two pawn islands: a/b-pawns and e/f/g/h-pawns.

2. Pawn Majorities - Identification

White has a three-to-two pawn majority on the queenside. Black has a four-to-three pawn majority on the kingside.

3. Space - Identification

Here's a comparison of each side's space along each file, using the method of counting squares that was explained in the prior articles of this series.

4. Holes - Identification

White has holes at e3, e4, and g3. Black has holes at b6, d5, d6, f6, and h6.


Now let's use these aspects of the pawn structure to help identify where each side's pieces, and possibly pawns, might belong. Because middlegame planning is affected by where the kings are located (and because the middlegame rarely permits relocating one's king), here's a diagram with the kings added.

1. Pawn Islands - Exploitation

Although Black's pawn structure is compact, the less advanced of his queenside pawns might still require protection by a piece. For example, so long as his b-pawn remains at b7, that pawn and his f-pawn would be hit simultaneously if White's queen were at b3 or f3.

None of White's queenside pawns is protected by another pawn. This magnifies their vulnerability and that of White's isolated e-pawn. For example, White's b2- and e3-pawns would be hit simultaneously if Black's queen were to reach the b6-square. So long as White is using only pieces to defend his queenside pawns, this will constrain the mobility of his pieces.

2. Pawn Majorities - Exploitation

Either side's attempt to use his pawn majority in the middlegame would be problematic due to the need to also prevent the opponent from invading along the open d-file. For Black, a middlegame advance of his kingside pawns would also expose his king.

3. Space - Exploitation

White has a pronounced space advantage along the f-file, and Black enjoys a comparable situation along the c-file. This suggests that White should attack the f-pawn and that Black should exert pressure along the c- and d-files.

4. Holes - Exploitation

White could try to exploit the hole at b6 by playing a5..., so that ...b5 or ...b6 could be met by an exchange of pawns and the transformation of White's c-pawn into a passed pawn. If Black refrained from advancing the b-pawn, then White might try to eventually plant a piece at b6 (although this would probably be difficult to achieve, given White's space disadvantage along the c-file).

The hole at d5 would seem an ideal post for White's knight, as it would be centralized, would prevent Black from invading along the d-file, and could be protected by White's c-pawn, e-pawn, or both. Also, a d5-knight would control Black's holes at b6 and f6 while being immune to exchange by Black's bishop (which is on the other shade).

The weakness of the holes at f6 and h6 is mitigated by Black's bishop's being on their same shade and therefore able to control those squares.

White's hole at g3 could be fatal if Black could play ...e4 and align his queen and bishop along the h2/b8 diagonal, because the move g3... (to block the diagonal) might well lose the pawn.


The above considerations indicate that White should post his knight at d5 and his major pieces along the f-file, as in the following diagram.

Next is how to place White's queenside pawns so that only one of them might require protection by a piece. One possibility is to put them at a4/b3/c4. In that case, the most convenient way to defend the b3-pawn would seem to be along the rank, obliging White to play e4... This configuration is shown in the following diagram. One of its drawbacks is that White's queenside pawn formation might be undermined if Black were able to play ...b5.

An alternative configuration for White's queenside pawns is a5/b4/c3. The a5-pawn would inhibit the advance of Black's b-pawn, especially with the b6-square's being controlled also by White's knight. The base of White's queenside pawn chain, namely the c-pawn, would be protected by White's knight and not easily attacked by Black's bishop (to put it mildly). Although the move e4... would not be necessary to open White's third rank, it might be desirable anyway to further protect White's knight, to obstruct Black's e-pawn, and to restrain Black's f-pawn. This configuration is shown in the following diagram. One of its drawbacks is the creation of a hole at the c4-square, although it's not clear how Black might exploit this.


Let's now return to the game position and consider Black's possibilities.

Black's rook and minor piece are far more mobile than their White counterparts. But if White were given time for 25. Nc3... and 26. Nd5..., this might allow him to stabilize the position. Therefore, Black's top priority is to prevent this maneuver, and its being Black's move gives him just enough time to control the d5-square with each of his major pieces. The suggests moving Black's rook to d8 and Black's queen to a square from which it controls d5, not necessarily in that order.

What about Black's bishop? One possibility is to attack the e-pawn by playing ...Bc5 or ...Bg5, the latter which would also pin the e-pawn against White's rook. Another is to play ...Bd6 or ...Bf6 after having played ...e4 to open the h2/b8 or a1/h8 diagonal, respectively.

Mentioned earlier was ...Qb6, forking White's b- and e-pawns. However, the reply 25. Qd7... would be disconcerting. This indicates that Black must maintain control of the d7-square.

Black can control the d5-square, while maintaining control of the d7-square, by either 24...Rd8 or 24...Qc6. 24...Rd8 would coax 25. Qf3, where it was noted the queen stands well. However, on 24...Qc6, 25. Qf3 would lose a pawn after 25...Qxf3 26. gxf3 Bg5, exploiting the pin against White's e-pawn and the pressure against White's c-pawn.

After 24...Qc6 25. e4 (hoping for 25...Qxe4 26. Qd7), 25...Bg5 would win the exchange (26. Nd2? Rd8). White instead played 25. Nc3, but 25...Rd8 prevented 26. Nd5. Black's continued harassment gave White no time to reorganize his pieces and soon culminated in a decisive gain of material. For the continuation, see "Erdai-Levin" under "Annotated Games" at this website.

In general, here are some reasons for moving a particular piece before others: (1) it would constrain the opponent (possibly by creating a tactical threat), (2) its best square is more certain than those for other pieces, (3) it can more readily be redeployed elsewhere in an emergency than can other pieces, or (4) the other pieces might cease to constrain the opponent if moved right away.


I hope that this discussion has conveyed a way of thinking that will help the reader determine where pieces belong and in what sequence they should be moved.