csipgs Chess

In comp.sys.ibm.pc.games.strategic (also known as csipgs), people discuss computer games they refer to as MoO, CIV, CotNW, CaveWars, and such.

No doubt you have wondered what it would be like if there were a Chess variant whose rules were patterned after these games. Wonder no more, the mystery is ended, and no longer need you lie awake at night in puzzlement.

The Rules of csipgs Chess

The rules of FIDE Chess apply except as follows.
Each player has a position on the board, a set of pieces in reserve, a treasury, and 6 piece designs.
In the initial board position, there is only a White King on e1 and a Black King on e8; initially, there are no pieces in reserve and the treasuries are zero; and at the start of the game the 6 piece designs are the FIDE Pawn, Rook, Knight, Bishop, King, and Queen.
Each turn consists of 3 phases: credit, action, and design. These phases must always happen in this order: credit, action, design.
1. Credit: Each turn, before anything else happens, one zorkmid is added to a player's treasury.
2. Action: Each turn, a player may either move a piece on the board, transfer a piece from reserve to an empty square adjacent to a friendly King, or buy a new piece for reserves. One and only one of these three actions may be taken per turn. If you can't do any of these things, you lose.
3. Design: Each turn, a player may also change one piece design. There is a discussion of this later.
There are a few restrictions on buying new pieces: you must have enough zorkmids in your treasury, you may not have more than 16 pieces (and this includes both your reserve and your position on the board, so if you have 8 pieces in reserve and 8 on the board, you can't buy a new piece right now), and you may not be in check.
You may not transfer pieces from reserve to the board using a King (or other royal piece) that is in check. Of course, you may not transfer if there are no empty squares next to your King, and you may not transfer if you have no pieces in reserve; and you may transfer only from reserves to the board, never from the board to the reserves.
If you have more than one royal piece, you may choose to ignore check; however, while you are in check you may not buy new pieces, and the "King" that is in check cannot be used to transfer pieces from reserve. (If you have only one royal piece, of course you must get out of check and of course the game ends if you are checkmated.)
Restated for emphasis, if you have two Kings and one is in check, you can ignore the check and use the King that is not in check to bring in a new piece from your reserves (but you can't buy any new reserves). If you have two Kings and both are in check, you cannot bring in any reserves.
Each player always has the right to know the size of the other player's treasury, the contents of the other player's reserves, and the other player's list of current designs. "Chess is a game of perfect information", they say -- although sometimes I wish I had more information about what nasty surprise my opponent is planning!
There is no such thing as Pawn promotion in this game, Pawns do not get an initial double step, and of course Castling is impossible.

Designing New Pieces

You always have 6 piece designs available; when you buy a new piece, you can only buy something for which you have a design. In other words, you can't design a new piece and buy it on the same move.

In order to design a new piece, you must scrap one old design. This has no effect on existing pieces of that type, it only means that you can't build any more of that piece until you redesign it.

Restating the previous paragraph more completely: suppose you want to design a new piece, and you choose to get rid of your Pawn-design and replace it with a Yawn-design. This has no effect on the Pawns that you already have in reserve or on the board, but it does mean that you won't be able to buy any new Pawns.

The effect of this rule is that you know what your opponent can place on the board two moves from now (but of course he has to stand still on the board in order to do so, therefore you can often calculate as far ahead as you do in FIDE Chess).

In order to design a piece, simply choose from the list of powers and modifiers. Add up the costs of the powers and multiply by the modifiers as indicated, round to the next highest whole number, and you have the price of the piece.

For example, suppose you choose Rook power, with the modifier "forward only". The cost of Rook is 5, the modifier is 0.5, that gives you 2.5, rounded up to 3 zorkmids. Now you decide that this is no good because you are wasting half a zorkmid, so you add the power of a Ferz to this piece, and you get FfR == 4.0 zorkmids.

You have now designed a new chess piece and derived an appropriate approximate cost (the prices in this game need not be as exact as they are for Chess with Different Armies).

In order to understand how to derive the symbol FfR for this piece, please consult my funny notation.

List of Powers and Modifiers

In order to understand the "atomic powers" referred to here, please consult Articles about Pieces.

Orthogonal Atoms

Base cost 1.5: Wazir, Dabaaba, "H" (the 0,3 jumper).

Base cost 5.0: Rook.

Modifiers for these are:

Forward-only, sideways only (multiply base cost by 0.5 for either one)
retreat only (multiply base cost by 0.2).
May use this power to move but not to capture (cost 0.6)
May use this power to capture but not to move without capturing (cost 0.6)

Notice that if you build a Rook from its component parts it costs more than buying a whole Rook all at once.

Other Atoms

Base cost 1.5: Ferz, Alfil, "G" (3,3 jumper).

Base cost 3.0: Knight (special modifiers are narrow and wide, each 0.5).

Base cost 3.3: Bishop, Long Knight (1,3 jumper, can be narrow or wide).

Base cost 5.5: KnightRider (can be narrow or wide).

Modifiers for these atoms:

Wide or Narrow (apply to Knight and Long Knight only), 0.5
Forward, 0.7
Retreat-only, 0.4
move but not capture, 0.6
capture only, 0.6

Notice that you can have a forward-only narrow Knight (from e4 to d6 or f6), whose cost is base 3 times modifier "narrow" 0.5 times modifier "forward" 0.7, giving 1.05, rounded up to 2.0; this piece appears in Shogi. (Because rounding-off made you waste 0.95 zorkmids, you'd want to add some more power to this piece.)

Whole-Piece Modifiers

Colorbound: If the piece is restricted to squares of one color, multiply its total cost by 0.9 -- you get a ten per cent discount for colorbound pieces, and a Bishop therefore costs 3 zorkmids.
Royalty: If a piece is royal, multiply its cost by 4. Note that if you have more than one royal piece, you have more than one place to bring new pieces from reserves, and also the rules of check are slightly changed.
Relay: This modifier is not officially available. If you agree with your opponent to allow it, the suggested modifier cost is 2.0, and please note that making a royal relay piece is risky.

Interesting but Unimplemented Rules

Plunder

Every time you capture a piece, the enemy treasury loses some fraction of the value of that piece. This is not a rule, at least not yet.

This rule would be intended to make it possible to get the opponent in debt and therefore prevent her from buying new pieces (for a while). This is why, instead of the usual rule (common in many csipgs games) of giving you a percentage of the value of the piece that was captured, the amount is subtracted from the other guy's treasury.

It is not yet known whether such a rule would be either good or necessary.

Notes on Game Design

This would turn out to be stupid and endless if you could just keep on cranking out new pieces and throwing them onto the board.

In order to prevent this failure of the design, I came up with the two step buying process (in order to get a new piece, you must waste two tempi, that is, you must go two moves without moving on the board). It also seems to me that the limited current designs and the reserves allow you to calculate better -- imagine what the game would be like if either player could simply make up a new piece and toss it on the board right away -- how could you plan ahead?

The restrictions on purchase and placement are there for the same reason. It's harder to defend against an attack if you can't interpose a paratrooper to block a check.

In addition, the fact that it costs you two moves to get a new piece onto the board seems to create strategic tension.

The list of powers and modifiers appears to be short, but in fact you can have thousands of different pieces in this game.

The rough-and-ready pricing scheme is far from perfect. You can lose by designing and building pieces that cost more than they are really worth.

Because the pricing does not need to be perfect, it is possible to include some atoms and modifiers that cannot be used in a set-piece game with predefined armies. The list should be even bigger than it is.

The rules about check with multiple Kings are interesting. If you build a forward-only move-only G, to use as an advanced base for new troops, it costs only 2 zorkmids. When you move it forward to serve as a source of trouble for the opponent, suddenly he attacks it! Now you cannot buy new reserves. Unless you can relieve the check or force the opponent to capture this "King", you are in big trouble.

There is no rule allowing pieces to capture themseleves. There are at least two situations in this game where it would be useful to be able to get rid of one of your own pieces, but you cannot do so.

Notes on Game Play

It is probably wise to have at least one piece in reserve at all times, and it is probably wise to have less than 16 pieces in play so that you can design and buy a new piece as the occasion demands.

It is probably wise to have all 16 pieces on the board, where they can move around and destroy the enemy.

It is probably wise to save up your zorkmids and buy a big, expensive, powerful piece. It is probably wise to have lots of cheaper pieces.

It is probably wise to buy as many pieces as soon as possible; later on you'll be too busy.

It is certainly a smart idea to be aware of the strategic tension between the several possible approaches to the problem of buying and placing pieces.

The Computer Challenge

Computer programs cannot play this game competently.

csipgs (moderated) Chess

The basic game of csipgs Chess can be played by two players, but it lacks one feature often found in the csipgs games it emulates.

The feature it lacks is a large map with special features that needs to be explored. In other words, you don't get to find out that b2.a1 is an impassable square until you get one of your pieces close enough to see that square.

In addition, you don't know where your opponent's pieces are until you get close enough to see.

In order to play csipgs chess under such conditions, you'd need a computer program to act as moderator between the two players.

Credits

The piece design system of csipgs Chess is lifted bodily from MoO2, one of the 5 best csipgs games ever (the other 5 are MoO1, CIV1, CotNW, and CaveWars, in no particular order).

Note added by Hans Bodlaender

Comments

csipgs Chess. Design and buy new chess pieces during play. (8x8, Cells: 64)

Jerdle wrote on Tue, Dec 18, 2018 03:39 PM EST:

The royal fmG isn't 2 zorkmids at all, it's actually 3.

So it's no cheaper than a royal fmGbmD, which fixes the problem of just having a royal fmG.

sirius628 wrote on Thu, Aug 31, 2017 07:34 PM EDT:

I would consider adding a (3,2) leaper as well as a lame modifier.

Jeremy wrote on Thu, Mar 24, 2011 11:43 PM EDT:

'Computer programs cannot play this game competently.'

What makes you say that?  If I were going to design an AI for this game,
I'd probably:
1.  Start with something like a FIDE Chess AI
2.  Modify the heuristic for the value of a board position.  Estimate piece
value based on cost, reserve pieces about a tempo less, bank may be
somewhat nonlinear but this doesn't seem scary.
3.  Create a representative library of about 20-50 useful pieces, including
at least 1 using each atom.
4.  For lookahead, you know the pieces purchasable on the current turn;
assume pieces purchasable on future turns are that player's current
designs plus entire library (limited by bank).
5.  Collect some stats from the lookahead to predict the designs you'll
need in the immediate future (AI never builds designs except those in
library).

Branching may be higher than FIDE, enough to weaken the AI, but I would
think it could still play respectably.  You can sandbag the AI if you
invent a piece that is significantly better for the current position than
any piece in its library, but that seems hard, and AI starts considering it
as soon as you design it, which is at least 2 turns before it can attack. 
Am I missing something?

Anonymous wrote on Thu, Jul 16, 2009 04:31 PM EDT:Good ★★★★

This is great, but I'd advise renaming the currency. Zorkmids were made for Zork, which has precious little to do with chess.

Anonymous wrote on Thu, Oct 21, 2004 03:10 PM EDT:

What would the cost of an immobile piece be? An example of one would be a Flag, which is immobile but royal.

Number of ratings: 2, Average rating: Excellent, Number of comments: 8