Hannibal Chess

This 10x8 chess variant was inspired by the larger (12x8) old chess variant Courier-Spiel. It occured to me that 'less might be more', if I reduced that game's board size, and dropped its two pieces that had the same short range characteristics as the guard and centaur piece types. In addition, the (Modern) Elephant piece type might prove more effective, given my chosen smaller 10x8 board size. This game's Elephant is also an old-fashioned third minor piece type that would remain as a possible attraction over standard (8x8) chess, and this apparently without it often taking players too many moves longer to get castled, or to get their rooks connected, as may happen in Courier-Spiel.

Game Courier presets for play are available, including a rules enforcing one by Jean-Louis Cazaux.

Setup

Pieces

There are 7 piece types in Hannibal Chess, 6 being the same as for standard chess (pawn, knight, bishop, rook, queen and king). The other piece type is the Elephant:

The Elephant is a compound piece that moves like a ferz (one square diagonally) or an alfil (i.e. leaps 2 squares diagonally). Note that this piece type is colour-bound, and is also more formally known as a modern elephant (or ferfil).

Rules

Castling is as in standard chess, except a king goes three squares sideways during the process (instead of two squares). Threefold repetition of position or stalemate is a draw as in chess, and similarly the 50 move rule is also in effect in Hannibal Chess (i.e. game drawn if no captures or pawn moves before 50 consecutive moves by both sides). Pawns move as in chess, and can promote to any piece type included in the setup, except for a king.

Notes

My tentative estimates for the piece values in Hannibal Chess are: P=1; (Modern)E=2.695(or 2.75 approx.); N=3.38(or 3.5 approx.); B=3.75; R=5.5; Q=B+R+P=10.25 and the fighting value of a K=3.2 (though naturally it cannot be traded). Note that some estimate a N (or B) as high as 3.5 on an 8x8 board, which I tend to agree with (on that board size I put [Modern]E=3.125).

Note that Dr. H.G. Muller finds that on an 8x8 board, N=B=3.25, and that a pair of Ns >= a pair of (Modern)Es (i.e. a pair are about the same in value), and he speculates that a lone (Modern)E might be about 0.2 less than a lone N (he also thinks that going to a 10x8 board won't change this relative difference in value much; his computer studies put N=3 and lone B=3.5 on that board size). Note also that in 2001, Zillions of Games barely preferred a (Modern)E over a N on an 8x8 board, by only about 0.06 of a pawn.

For those who might wish to have a physical 10x8 chessboard, Dr. Muller has mentioned elsewhere that a simple solution would be to cover up two ranks of a physical 10x10 chessboard. For those who might wish to search for a more elegant solution, at least in the past such boards were sold in Germany for the 10x8 variant Janus Chess (the pieces of its set could be somewhat suitable to play Hannibal Chess with, too). Here's a link about that game, with the specifics for ordering a set included:

Janus Chess

Note that ChessV currently supports Hannibal Chess.

Finally, a variant idea of mine some may wish to try would be Lieutenant Chess. This would use the same rules and setup as for Hannibal Chess, except to replace each side's (Modern)Es with Lieutenants (from Spartan Chess). Lieutenants have the same powers as (Modern)Es, plus they can instead take a single-step non-capturing move sideways, i.e. for one square in either direction along their current rank, and thus they are not colour-bound pieces. On a 10x8 board, such as for this variant idea, I'd tentatively estimate a Lieutenant to have a value of 3.97(or 4 approx.). I'd note also that I don't recommend using FAD pieces (Betza's notation; also see CVP site's Piececlopedia), instead of (Modern)Es as in Hannibal Chess, even if using Fast Castling rules as in my Wide Chess CV invention, as they might often be awkward/harrassed it seems to me, but some may wish to experiment with such a Mutator idea (it could be called [10x8] FAD Chess).

Last revised by Kevin Pacey.