- Some start positions are too bizarre or illogical for serious play
GeneM introduced his example of a family of bizarre positions in two separate comments:
Almost half of the chess960 positions have a Bishop start on a corner square. Such setups are bad because the Bishop has only one way to develop. Those setups should be discarded. As Kramnik noted, a Bishop that starts on a corner square has only one degree of freedom in how it can be developed. It needlessly reduces the range for human imagination in using the pieces from the start of the game.
This reminded me of a concept I documented some time ago, where my first problem was to locate the post. I found it on my main blog, where I used to write about chess960 before setting up this current blog which is dedicated to the subject. While searching for the post, I realized that I had never incorporated my earliest posts about chess960 opening theory into this chess960 blog. Here they are, in reverse chronological order, because later posts tend to build on their predecessors.
- 2009-12-08: What Makes an Opening Extravagant?
- 2009-04-14: A Framework for Chess960 Opening Theory
- 2009-03-31: Introduction to Chess960 Geometry
- 2009-03-24: Comments on Chess960 Opening Theory
- 2009-03-17: Lasker's Table of Opening Values
- 2009-02-24: How Top Players Treat the Same Chess960 Position
- 2009-01-27: Chess960 Opening Theory
- 2009-01-13: Undefended Pawns in Chess960 Start Positions
- 2009-01-06: Advantage in Chess960 Start Positions Revisited
- 2008-12-30: A Followup, an Error, and an Insight
- 2008-12-16: Theoretical Advantage in Chess960 Start Positions
- 2008-12-09: A Database of Chess960 Start Positions
- 2008-12-02: An Exercise in Chess960 Positional Thinking
- 2008-11-11: Lesson in Chess960 Opening Patterns
- 2008-11-04: Chess960 Game Explorer
The post I was looking for is the second in the list: A Framework for Chess960 Opening Theory. The 'framework' is a two dimensional array of Pieces & Possible Start Files together with a notation to identify cells in the array. For example, GeneM's example of a 'Bishop that starts on a corner square' could be identified B:a/h, i.e. a Bishop starting on the a- or h-file.
Unlike GeneM & GM Kramnik, I enjoy playing B:a/h positions. Rather than taking two moves to develop the Bishop as with the B:c/f of traditional chess -- a Pawn move followed by a Bishop move -- the corner Bishop is developed by a single move: advancing the adjacent b-/g-Pawn on the diagonal. On top of that, the corner Bishop never interferes with castling, meaning that the two operations -- (1) Bishop development & (2) castling -- can be executed independently. In traditional chess, the one always precedes the other. Furthermore, the development of the Bishop sometimes uncovers an attack on a weak Pawn on the Bishop's diagonal. When this happens, tactical complications arise immediately.
I see no reason to single out the corner Bishop as 'bizarre or illogical'. A corner Knight (N:a/h) has limited options because its first developing move is usually to the b/g file rather than the c/f file. A corner Queen (Q:a/h) often means that the Queen is slow getting into the game, giving these positions a slow, positional buildup (see 'Fianchetto the Light Squared Bishop' for an example). Since a King can't start in the corner, the only piece really suited to the corner is the Rook (R:a/h), like in traditional chess. As with the Queen, a corner Rook has the disadvantage that the piece is slow to get into the game. Early Rook actions, which can be compared to a lightning tank attack in modern, mechanized warfare, almost never happen.
What about the Bishop starting on other squares? The Bishop starting next to the corner (B:b/g) also has limited options. Its natural diagonal, the one chosen for its development in most games, is the long diagonal, opened by moving the c/f Pawn, rather than the short diagonal, opened by moving the a/h Pawn. Starting two squares from the corner (B:c/f), is the setup we all know and love from traditional chess, while the central Bishop (B:d/e) offers a different set of challenges. I quoted GM Seirawan on this last possibility in 'A Tempo and a Half in a Symmetrical Position'.
The only real disadvantage of the B:a/h setup is when both Bishops start in the corner. When this happens, all four Bishops are facing each other on their long diagonals. The order in which the Bishops are developed becomes a subtle tactical dance where a player's fast grab of one diagonal cedes the other diagonal to the opponent. Furthermore, a premature development of the Bishops can lead to them all being swapped off in the opening for a Bishopless middlegame. A player who wants to avoid this must block the diagonal before developing the Bishop, but this gives the opponent the opportunity to develop first on the same diagonal. And so the dance continues.
One point which should never be forgotten: whatever the advantages and disadvantages of a specific start position, both players are struggling with the same issues. The only difference is that one of them starts first. This is a feature of traditional chess with which we have all learned to live and is no less true for chess960.
[Note to myself: Determine how many positions have B:a/h,b/g facing a weak Pawn on the diagonal. Ditto for Queens.]