14 July 2012

Non-random Fischer Random

In Kasparov's Modern Chess, I touched on the idea of Bronstein chess.
Re Sveshnikov's 'more sensible is Bronstein chess', I believe that he is referring to the variant of shuffle chess where the two players take turns placing the pieces on the back rank. I've never investigated this idea and it would make a good start point for a followup post.

This is confirmed in the Telegraph's 2006 obituary of David Bronstein.

He was one of the originators of Rapid Chess played at a faster time limit, with 30 minutes or less for the game, and developed a form of Random Chess well before Bobby Fischer claimed ownership of the concept. In Bronstein Random Chess the pawns are set out and the first eight moves involve placing the pieces on the vacant back rank.

HarryO pursued this idea in a comment to the Modern Chess post, where he proposed that the players alternate placing the King, Queen, etc. until all pieces have been placed. While this is certainly an easy method to follow, it has the drawback that it can favor certain of the 960 possible positions. We have already seen this problem in the method used to select start positions during last year's 'Kings and Queens' event in St.Louis; see the comment to The Chess960 Wheel of Fortune.

Since the only method to ensure an even chance of choosing all positions is to start by placing the Bishops, and since the position of the King and Rooks is determined by the start squares of the other five pieces, I propose a different method.

  • Player A places a Bishop, which the other player echoes.

  • Player B places a Bishop on a different color square than used in the previous step, also echoed.

This leaves the Queen and Knights to be placed. I propose that the two Knights be placed on the same step by the same player, and that this should be considered a single step.

  • Player A places either the Queen or the two Knights, which the other player echoes. The choice of placing either the Queen or the Knights is Player A's decision, as is the choice of square(s) for the piece(s).

  • Player B places the remaining piece(s) -- either the Queen or the two Knights -- also echoed.

  • The King and Rooks are placed on the three remaining empty squares according to the rules of chess960, the King between the Rooks.

This method ensures that all 960 positions have an equal chance of being selected. It also allows for the development of a new kind of opening theory. As for the question of which player gets to go first as Player A, that can be decided using the same sort of method that we already use to determine who plays White.

[To be consistent with the standards of this blog, I should have titled this post 'Non-random Chess960', but the play on words was too cute to ignore.]

2 comments:

HarryO said...

Hi Mark good idea to put the bishops down first! After all from what we have noticed is the placement of the bishops seems to be the greatest controversy!

Love the idea of then placing the knights down simultaneously considering they are uncoupled from square color.

So just to simplify a bit, how about this?:
1. Bishop down player A and player B must mirror.
2. Bishop down player A and player B must mirror.
2. Knights down player A and player B must mirror.
3. Queen down by player B and player A must mirror.

It would be great to trial the idea in a real world club please someone do that if you can!

Only people who have played Chess960 know how the psychology of the start position is determined by the structure of the minor pieces....

Hypothetical scenario between two players. One is Ms Aggressive and the other Mr Positional:

1. Ms Aggressive puts a bishop on b1 hinting that she is thinking of putting the next bishop down on a1 for a huge cross board attack. So she plants down the bishop as such:
xBxxxxxx

2. Mr Positional doesn't like the idea of a bishop down on a1, and so puts down a bishop on e1 throwing out the balance. He puts down the bishop as such:
xBxxBxxx

3. Ms Aggressive didn't like that much but knows she now can swing it back her way by putting down the two knights. Because she has played Chess960 extensively, she realizes that where the two knights go really changes the character. Ms Aggressive decides not to put the knights down on the same colour because she doesn't want to deal with the possibility that the knights will get shut down by Mr Positional because they are on the same colour. At the same time she doesn't want a knight in the corner because she wants activity! Therefore she puts down the knights as such:
xBNxBNxx

4. Mr Positional must mirror the knights, but now gets the chance to place the queen. Mr positional has played a lot of Chess960 and has learned that a queen in the corner is a really tricky "positional" scenario. At the same time, Mr Positional has learned that queenside castling is more problematic and wants to castle kingside quickly if possible. Therefore Mr Positional puts down the queen as such:
xBNxBNxQ

Now it is settled after much "psychological warfare" even before the game has begun! The start position will be:

RBNKBNRQ = SP568!

Ms Aggressive has gotten what she wants to some degree. At least the king is the center of the board. Mr Positional is happy because he still has the option of quickly playing out the knight and bishop and castling kingside with the queen tucked away in the corner to lesson tactical activity.

Let the game begin! Is this the future of Chess? The players "evolve" the start they want. Because the opening is unknown, the games take longer to play. A real contest from the moment the players arrive at an EMPTY BOARD.

The beauty of this idea is that it doesn't really matter if there is equal chance of all 960 positions appearing, because humans will "evolve" it themselves over time. At least Marks idea minimizes too much repetition of the position.

See the point is that players will gravitate to SP518 traditional chess but over the generations will move further and further away from it. If player A puts down a bishop on C1, this hints at SP518. The second player then must decide how far they want to deviate from SP518.

Interestingly, it will be BLACK that decides the deviation by the choice of the second bishop position....

This is ideal, because black has the power to freshen the position and feel like they have more of a chance to dictate the flavor of the start because they don't have the first move freedom that white has.

Cheers

HarryO said...

Correct myself in the above post!

How about this?:
1. White puts a bishop down and black must mirror
2. Black puts the other bishop down and white must mirror
2. White puts both knights down and black must mirror
3. Black puts the queen down and white must mirror

So the trade off for white picking the knight positions and having the first move, is that black picks the queen position which also dictates the final position of rooks and king.

Cheers
Harry