Re: Chess variant problem - a bug in Zillions?

John Rowland wrote:
> ...
> When Zillions plays my game against itself, white wins nearly all 
> of the time in various symmetrical starting positions. The obvious
> conclusions were that my game was flawed or that my rule file was
> flawed, but the test described below suggests otherwise.
> ...
> All of the pieces have symmetrical move possibilities, including 
> my "pawns" which can move (or capture) backwards as well as 
> forwards, so I have not needed the Zillions "symmetry" command 
> to make black's pawns move in the reverse direction.

Maybe the absence of symmetry commands is the reason for the
differences in the play of white and black.  

I assume that the move generation will be done in the order 
of the statements that you have coded in your zrf. 

If (for example) you write in your zrf for the "pawns"

  (moves
     (pawn-move n)
     (pawn-move s)
  )

then Zillions will generate with the white pawns:

  first the forwards (offensive) moves, and
  second the backwards (defensive) moves - 

but for the black pawns this order is the contrary: 
  defensive moves will be generated before the offensive moves.

Similarily this should happen with the other pieces too.


Friendly greetings,
Alfred Pfeiffer

P.S. It would be nice we  could seem your new game.