My incoming variant, Chess on a Larger Board with not-so-few pieces dropped, tries to graft on my favorite 12x8 pattern a variety of rather different pieces, like David Howe did in Chess on a Longer Board with a few Pieces Added, which features the Wall.
My two-square piece is a Golem (nothing to do with Golem Chess). When another of my pieces, the Wizard, is taken by a Pawn, the Pawn and the Wizard form a two-square diagonal piece, whose both parts, which must remain connected, move first on the symmetrical square on the same rank, which need not be empty, and then one square diagonally. (Thus the Golem's path is both side-changing and color-changing.)
A Golem has up to nine moves. For instance, a Golem on c2 and b3 (the files are indexed from y to j because the starting array on the eight central files is duplicated from Orthochess) may move as:
c2 to g3, and b3 to f2, f4, h2 or h4
c2 to g1, and b3 to f2 or h2
c2 to h3, and b3 to f2 or f4
c2 to h1, and b3 to f2.
The Golem captures by replacement, and is captured when either of its parts is captured. (Which comes fast, because it is created anywhere on the board, and is not as well guarded as the Wall, but otherwise the owner of the Wizard wouldn't allow the formation of an enemy Golem.)
Of course, the Golem could have been devised to move first on the symmetrical square on the same rank, and then one square orthogonally. There would still be nine moves, and the Golem's path would be colorbound.
(Actually, it is nearly colorbound. It has to cross the center files before reaching a square of the opposite color on the same side of the board.)
My two-square piece is a Golem (nothing to do with Golem Chess). When another of my pieces, the Wizard, is taken by a Pawn, the Pawn and the Wizard form a two-square diagonal piece, whose both parts, which must remain connected, move first on the symmetrical square on the same rank, which need not be empty, and then one square diagonally. (Thus the Golem's path is both side-changing and color-changing.)
A Golem has up to nine moves. For instance, a Golem on c2 and b3 (the files are indexed from y to j because the starting array on the eight central files is duplicated from Orthochess) may move as:
c2 to g3, and b3 to f2, f4, h2 or h4
c2 to g1, and b3 to f2 or h2
c2 to h3, and b3 to f2 or f4
c2 to h1, and b3 to f2.
The Golem captures by replacement, and is captured when either of its parts is captured. (Which comes fast, because it is created anywhere on the board, and is not as well guarded as the Wall, but otherwise the owner of the Wizard wouldn't allow the formation of an enemy Golem.)
Of course, the Golem could have been devised to move first on the symmetrical square on the same rank, and then one square orthogonally. There would still be nine moves, and the Golem's path would be colorbound. (Actually, it is nearly colorbound. It has to cross the center files before reaching a square of the opposite color on the same side of the board.)