A thought on pieces that move one way but capture in a different way

In chess variants, almost all pieces have a certain way of moving, and how they move is how they capture. Very few pieces have a way of capturing different from their way of moving. (Then I am for now putting aside the many pieces that have no capture at all.)

List of pieces that, to my knowledge, have a way of capturing other than their way of moving:
- FIDE pawn (not the pawns from Xiangqi and Shogi variants: they capture forward how they move.)
- FIDE King: castling is meant as a move without capture.
- Hunter: invented by @HorribleTomato: moves like a King, captures like a Knight or leaping orthogonally to the second square
- Xiangqi Cannon: moves without jumping, captures with one jump
- Great General, Vice General, Rook General, Bishop General from Tenjiku Shogi: they can only range-jump while capturing, not while moving to a vacant square

The idea is of course this:
From any two pieces A and B that move the same as they capture, we can derive two new pieces: One piece moving as A but capturing as B, another piece moving as B and capturing as A.

I foresee one difficulty: we'll soon be running out of names.

Nice comments on it. Great to see a list of the pieces. I've been trying to figure this out actually

Here are some of mine. I don't have fixed names for them
A piece that moves like a rook and captures by sliding away from the captured piece as many spaces as it would have moved if it had moved onto the captured piece instead
A piece that moves like a rook and captures by making a leap to an empty space over the captured piece where the captured piece is in the middle of the line between the destination space and the origin of the leap, any other piece on that line are leaped over without being affected but the destination must be empty
A piece that moves one step and captures by sliding any distance
A piece that moves by sliding any distance but captures with a single step

I'd like to suggest a naming convention. These pieces can be called a 'move-capture-compound' and anabbreviation for this can be 'MCC'. E.g. the Rook-Bishop-MCC is the piece moving like a Rook but capturing like a Bishop. The Bishop-Rook-MCC is the piece moving like a Bishop but capturing like a Rook.

So from the existing pieces Rook, Bishop, Knight, King, we can already derive 16 new pieces and there's no need to come up with 16 new names for them. What I would like to try is a game with two armies consisting entirely of these MCC's.

correction - only 12

Graphics for an MCC:

Bottom piece shows how this MCC can move, top how this MCC can capture.

Some more graphics:

Possible starting position for a game:

I'd play a game of that against someone

Some remarks: a piece that captures in another way as it moves is commonly known as a divergent piece. Names one encounters are often chimeric fusions of the component names, such as Keen or Quing.

I once measured the values of such pieces; in first approximation these turned out to be 2/3*capture + 1/3*non-capture components. E.g. a Quight (mQcN) is worth aboyt 5 Pawns, a Kneen (mNcQ) about 7. Of the King-Knight hybrids it was interesting that the Kning (mNcK) was significantly (about 0.5Pawn) stronger than the other combinations (all worth about 3). It benefits from the concentrated attack power of the (non-royal) King (endowing it with mating potential, as well as from the speed of the Knight.

So according to this, I 'invented' divergent pieces:

Rishop 3,66
Right 3,66
Rard 4,33
Book 4,33
Bight 3
Bard 3,66
Knook 4,33
Knishop 3
Knard 3,66
Guook 4,66
Guishop 3,33
Guight 3,33

I had a really fun game with the hunter against @HorribleTomato some time ago. Here's its move diagram. Game is (here) if anyone is interested in seeing it.

I'll play with the hunter. You up for a challenge?

I'll play - should be fun. I've been a bit busy lately so I might not make a move every day, but I'll follow the 3-day time control. Do you want to play rated, or non-rated? Either way is fine for me.

Rated I guess. Even if you can't make the time limit, it's not a tournament, so no one is in a rush to get the game over with. As long as you move when you can, that would be fine

Ok great. Game is rated, and you can play White. (Here's the game).

Thanks!

I received a question if the King can castle with a compound of a Rook and another piece. This is not the case in most other variants. E.g. in Capablanca chess, the King cannot castle with a chansellor. It makes sense to follow this convention with these divergent pieces.

Actually, the Witch gives two other examples of divergency.

- The Witch can move but cannot capture at all.

- The bewitching ability is King-like, because only adjacent pieces are made transparent.

One can think of a bewitching ability according to any piece's movement. Not that I would recommend a Witch that makes transparent according to her own movement.

Or looking at Superchess Magician or Zombie: the Zombie freezes the same way as it moves, but the Superchess Magician is divergent from this point of view.

I did some testplaying, and tend to conclude that

- Adding a few of these divergent pieces to another existing variant is exciting and gives energy to the game

- A game with entire armies of divergent pieces is troublesome, boring and drawish

The 1983 game "Loonybird" (https://www.mindsports.nl/index.php/the-pit/537-loonybird) (https://www.chessvariants.com/ms.dir/loonybird.html) contains all 6 combinations of capture/move derived from Rook/Bishop/Knight powers. What I *haven't* seen is splitting the power THREE ways rather than two. Each piece wound move one way, capture another *and check the king a third*. The three basic powers of Rook/Bishop/kNight give six hybrid pieces -- RBN, RNB, NBR, NRB, BNR and BRN -- where the first letter is movement, 2nd is capture, third is checking power.

Checking power is a special case of a capture power. First we have pieces that can capture, then we have a King that must at all cost be saved from capture. In that context we defined check, mate and stalemate.

OK, we might step away from that context, and define checking power as something in itself. I don't see the elegance. But still an interesting thought.