Pond Scum

Pond Scum: a 10K variant

This is an attempt to push chess way beyond the boundaries of any previous variant*. It is played on a 100x100 square board, hence the "10K". It is not meant to be a theoretical exercise in nice-sounding ideas, but an actual game, playable, interesting, and fun. Otherwise, why bother?**

*Okay, after thinking a little, I remembered 2 variants that are in this size range, amazingly enough. One is David Howe's Mega-Chess, with 4096 [64x64] squares, and the other is Dale Holmes' Salmon P. Chess, at 7500 squares.
**The 2 examples above are 2 great reasons to bother, even if the games were never played. David's idea is a recursive classic, and Dale's game is just awesome. So I'll settle for a game that's just a little bigger, and hope it's not just awful.

Where does one start a project like this? Ten thousand squares is a lot of territory, easy to get lost in. A guide would be nice. In keeping with the fine tradition I started a month or two ago, the 24x24 square Fortress Chess, built upon the lessons of a successful 12x16 game, Chieftain Chess, will be the guide, as there is nothing nearer in size that I know of which fits the bill. While Chieftain Chess has a rules page and public preset, Fortress Chess is still experimental. Its rules are laid out in The Chess Variants Pages Comments Section under the topic "Big Board CVs", and the preset was still private, but no more! The URL for the Fortress preset is:

The key thing Fort showed is that using the concepts of leadership and command control will give structure and organization to a really huge chess army. The idea of combining simple piece icons with simple movement-indicator icons is also very useful. Finally, it acts as a caution about the proliferation of piece types and numbers. I'll be relying on these ideas to keep the armies manageable, but must obviously come up with something new to make the game [humanly] playable.

Being lazy, I'll copy something I wrote in the CV Comments on 2007-03-09:
"[T]he formations of wazirs and stewards are the forerunners of a new type of 'piece', consisting of several mostly shortrange pieces and a leader unit specific to them that they must be in 'contact' with to move. These would be 'Autonomous Multiple Pieces', or AMPs. While the 2 examples I've discussed so far are simple and slow, if these amps evolve a bit [a 3rd piece would be 6 forward-only ferzes and their leader - to make it a better attack piece, up the number of its components allowed to move each turn], their natural habitat would likely be on boards of side 30-50. I see them evolving specific organs [pieces] for attack, defense, and movement. But they are for later, larger games. I'd call those variants 'Amoeba Chess', but that name is taken by a game [by Jim Aikin; preset by A. Sibahi] that has a board that changes shape slowly, so maybe I'll go with something like 'Puddle Chess', where 2 groups of 1-celled critters fight it out for control of a splash of water on a city sidewalk."

Autonomous Multiple [or Multi-Unit] Pieces:

An AMP is a closely-spaced group of individual chesspieces that includes one or more leader pieces. Any individual chesspiece in an AMP must be "activated" by a leader before it can move. A simple AMP could consist of 4 wazirs and a guard unit acting as the leader, or 2 guard-leaders and 6 forward-only ferzes. More complicated AMPs would include a few leaders of various abilities and a range of chesspieces, for a total of a couple dozen units in the AMP.

Command Control

Chesspieces, except leaders, may not move under their own power. They must be activated by a leader to move.

* No chesspiece may ever move unless it is activated by a leader which has to be within a specified, small number of squares of it at the start of its move. The leader that activates it may have been activated and moved to that spot earlier in the same turn.
* An activated chesspiece may move outside the activation range of the leader which activated it, or any other [friendly] leader.
* Each leader may activate 1 or more chesspieces per turn, depending on the level of the leader. It may activate itself or other leaders.
* No chesspiece may be activated or move more than once per turn.
* Once a chesspiece has finished its move, it becomes inactive again. It cannot move in a subsequent turn without being re-activated by a leader.
* Leaders may activate chesspieces in their own AMP, or capture leaderless chesspieces and add them to their AMP. Under certain circumstances, a higher-level leader may capture a lower-level leader.


Chesimals are chess "animals". They are simple to complex AMPs with leaders that can activate a fairly large number of pieces per turn.

Their origins lay in a number of areas: military history and wargames; Conway's "Game of Life"; Valentino Braitenburg's excellent book "Vehicles - Experiments in Synthetic Psychology"; numerous biology texts, and essays by Stephen Jay Gould, Loren Eisley, Konrad Lorenz and others.

The simplest forms consist of chesspieces that all move only 1 square, wazirs, ferzs, guard-leaders. But these are the pawns of our game. More advanced critters have a "skin" made up of the simplest chesspieces acting as protection for the chesimal, and also have various "organs" which serve as weapons of offense, chesspieces with a little range. And I'll admit here that I am quite curious to see what the higher forms will evolve into.

Progress Report I

The first simple pieces are evolving before our eyes. We have to define a new term: "Level". The "Level" of a chesimal is the number of different kinds of non-leader pieces that chesimal has. Levels 1 and 2 chesimals have appeared in games. These have been simple, very short range pieces. Links follow:

The next steps in the development of our pond scum involve developing Level 3 chesimals and faster, more complex Level 2 pieces. Our test level 3 ortho amp has 4 1-step units, 4 2-step units, and 2 [3,0/3,1] leapers that may also move instead as ferz, with a bent sliding general leader. Our test upgraded Level 2 ortho amp has 4 2-step units [the standard D/W piece] and 4 3-step units [the D+W piece - see "Joe's Strange Notation"] along with its brain unit. We are now developing chesimals with attack and defense skills.

The first faint stirrings of life are being felt in our puddle. If we could not people our 10k board directly with pieces, we certainly seem to be on track to evolve them, which seems far more appropriate for pieces that are going to be creatures in a game called "Pond Scum".

It would be nice to have an arena where players could fight different chesimals to get a feel for how they work. Maybe someday.

Discussion 1/9/08

Is there a third way to flexibility, one that combines the properties of both chiefs and chesimals?

While a decent game in itself, Chieftain Chess is "too perfect" in its patterns and too chaotic for higher-level extrapolation. Played with the Symmetry rule, the game is crystalline in its individual patterns, but while a sequence of moves can be beautiful and flowing, it's a choreography of perfection, and is not easily manually reproduced at the large scale - at least, I certainly wouldn't want to move a dozen or maybe more even pieces in a perfect pattern each turn. And without the symmetry rule, the game can become pretty messy.

Chesimal Fusion I turns into a game of "Blob vs Blob", which is bizarre and fun [well, I like it, anyhow], but probably not always chess at its finest. Also, not real extrapolatable to a much larger size. You'd get megablobs of pieces in the 50-100 square sizes, and it'd be a real pain to visualize a strategy, and even one to move them properly by hand. Watching a few such blobs locked in combat and eating each other might be interesting a few times, but it's a *lot* of work to do it right.

Chesimals I more closely approaches my original idea of what would work at large scale, but it is a bit cumbersome to actually move, or verify the move of, AMPS that consist of several different kinds of pieces that all must remain in contact to move, pieces that would be moving roughly 10 squares a turn and would approach 12-15 individual "units" or more in size. It needs playtesting still, but, most oddly, people have not jumped at the chance to move groups of about 8 chesspieces around a 12x16 board.

So, combine the chief with a ranged command control [of 3 squares] for 1 piece, and the chesimal brain cell with its need to be physically contiguous with each chesspiece unit to move and it's ability to move about 5-10 units. A halfway point, or near enough to start, would be an "open chesimal", with maybe 5 units total, the brain and 3-4 others. The individual units of the open chesimal [need a better name here], along with the brain unit, must be touching or each up to 1 square apart. This means a 5 unit open AMP could form a line with the 5 chesspieces alternating with 4 empty squares, and all the units would be in contact with the brain unit.

There are two problems here, the problem of range, and the problem of optimal size of an "open AMP". A little bit of "fine maneuverability" might be thrown in to enhance a piece's attack effectiveness. Let's try to build some pieces. [I say "lets" here, implying you will be helping, because I have very little idea of where this idea will go, and could use the help!]

I think I will start with "daisies". These are multi-step linear hoppers, until their last step: Working the DW and AF [and maybe the DF and AW] to death.


"A daisy is a stalked plant. Maybe the stalk is short, and maybe the stalk is long, but the stalk always grows first, then the flower can bloom at the end of the stalk.
Stalks are Dabbabahs and Alfils; they can grow any length.
Flowers are Wazirs and Ferzes, they only happen at the end of the stalk.
You may have a stalk length of zero squares, giving the pieces just the wazir or ferz move.
A stalk length of two squares gives the hero and shaman.
This piece might be nice for breaking into defended areas as its range increases.
The movement track is a saw-toothed pattern."

David Paulowich and I were discussing a piece of this idea a while back, a 3-step dababbah-rider [a linear piece], and he was the first to suggest in print that the final step could be bent, so I promised I'd give him this much-deserved mention. I'd generalize it to "the last leap made can be bent", and add it to my rules mix here, for sure.

You've got some options in combining daisies with the Paulowich bend.

  • you can keep the 2 piece-types separate
  • you can allow one piece both options, but not on the same turn
  • you can allow a piece both options on the same turn - you might consider limiting the total range of this really nasty move.

1/10/08 [cont from above]

Let's start with 2 pieces, a "matched pair" of DW and AF, along with their leaders.

The DW7 WarMachine -
This piece may move in a straight line as a dababbah 7 times;
or may move 6 times as a dababbah and once as a wazir, in a straight line.
It may move 5 times in a straight line as a dababbah, then once as a dababbah or once as a wazir, in any direction.
It may move 4 times in a straight line as a dababbah, then once as a dababbah and once as a wazir, in any [1 or 2] directions.

Another option would allow the piece to choose a wazir move at any [or every] step of its move.
Clearly, you could alter movement specs - just for example, anyone would be able to figure out what a DW6 or DW8 was, and how it worked.

The AF7 WarElephant is the diagonal equivalent of the orthogonal WarMachine, jumping 2 squares diagonally as an alfil or stepping 1 square diagonally as a ferz, and getting "the Bends" with the same costs and restrictions as the warmachine. The move as ferz any or every step is also the same.

The DWAF7 General is the leader unit. It also moves up to 7 steps, as either the WarMachine or WarElephant, above, with these exceptions.
It has no movement penalty for making its 2 "bent" moves, it may always move up to 7 steps.
It may, in place of a standard "bent" step, switch movement modes: if it was moving as a DW, it may move once [or twice] as an AF, and vice-versa.

DINOSAURS, the game?

I see large, slow-moving wazir AMPs, slightly smaller ferz AMPs, and some few "XYn" series of pieces, but it needs something else, some medium-range pieces that are different. And some packs of small, fast predators.
One logical piece to add would be the knightrider, with either a wazir or ferz move at end, but maybe some additional "bend" restrictions.
A knightrider with a king move at the end, interesting piece. Possibly the piece would be allowed some patterns of knight moves for flexibility.


** 3/17/08 **
It's amazing how you plan one thing that doesn't work all that well, and something else suddenly strikes you as an alternative. At this point, Dinosaurs seem to be in decline - they may well make a comeback later, but right now, a simple form of "military chess" has scaled to board sizes in the 250-500 square range. But as well as growing, Chieftain Chess [aka: Gas Hogs] has shrunk down to 8x12 and 10x10 sizes, with 16 pieces/side, for accessibility. And I've learned a bit about extremely large games.

Structured Multi-Move
To play a big game in a reasonable number of turns, you allow each player to make more than one move in a turn. When the games get really big, and you're moving 5-10 or more pieces per turn each, things can get pretty chaotic unless you have a form of control. By providing activation rules, where pieces must meet some sort of extra condition beyond "it's my turn" to move in any given turn, you can impose a certain amount of order and regularity to the game. The method being used to impose order here is based, appropriately enough, on a military metaphor. The pieces are imagined to be combat units, whether they're seen as giant exhaust-belching robots in a post-apocalyptic world, or little half-naked people with pointy sticks running around on a sun-drenched plain.

The field of civilian wargaming has advanced considerably from 50 years ago, when The Avalon Hill Company was forming. Chess is often considered the original wargame, but has gone through a serious evolution since its inception, that has moved it away it from its origins. The series of military chess variants presented here have abstracted several things from modern board wargaming.

The key ideas abstracted from wargames were:

  • large board with low piece density;
  • unit standardization;
  • shortrange piece movements where pieces move only a fraction of board length in a turn;
  • multi-move turns;
  • the replacement of 1 king with several leaders;
  • command control, or the requirement that, for a piece to move, it must be within a specified number of squares of a special piece at the beginning of its move.

The game is playable at larger sizes because it can easily be "chunked"; that is, broken into smaller, more manageable pieces.
One reason for this ability to chunk the game is that the armies are made up of repeating groups of 8 units: 1 Leader, 1 Hero, 1 Shaman, 1 Skirmisher [aka: Scout], and 4 warriors, representing only 5 different types of pieces.
This 8 unit group containing 1 leader forms the basis for scaling the game. Such a group sets up on one side of a 12x4 board section, and a similar group is set up for the opponent on the other side of that 12x4 board section. So a multi-player game [which this series is well-adapted to] with N players on one [if uneven, the larger] side would use a board of size 12x4N - actually 12x4NL, where L is the number of leaders per player on one [larger?] side.

The following is from an email to David Paulowich

The draw question bothers me greatly in this whole series of games. 1 guard plus chiefs vs chiefs should win - if you can pin the chiefs against the side and send the guard in as an initial sacrifice. I don't know if there's a board size where it's impossible, but I suspect there are many boards where it's "difficult". Even 12x16 with 1 chief vs 1 chief plus 1 guard is problematic… now to your strange question.

STRANGE QUESTION: can a player draw the game by forming his army into a 5x5 block and just defending?

Eric Greenwood started to do just that before he lost access to the CV site from his library computer. It started to worry me. I do think the mobile player can pick spots to attack, nibble at corners, try to use the block's lack of freedom of motion in the middle against it, overload to one corner… still don't know if that will work, but having the initiative in a wargame is worth something. [Even if it only means you don't get killed until the end of the turn, instead of the beginning. I've played games where I had a chance to run futilely.]

On 20x30 - you now see my elephant as moving 1, 2, or 3. How about 4? Or 6, 8, or 10? I'm thinking about leaders that can move 3-6 pieces, all within 2 squares, no contact necessary, or within 2 squares of a piece within 2…
What about a leader that just transports 3-6 pieces up to 10 squares in any direction, with the pieces maintaining the same relative positions, capturing on the landing space[s]. When a unit [as opposed to a multi-unit "piece"] in transit lands on an enemy unit, the enemy unit is captured, and the friendly unit stops moving, although the friendly piece [AMP] keeps moving. On any given turn, a leader may activate shortrange units to move with their individual shortrange moves, or carry units x number of squares, or activate some and carry others. A unit may either activate or be transported in a single turn, not both. Transporters.

These AMPs will need room 20x30 is barely enough. "Piece" density is very low, and even unit density can be quite low 5 - 10 - 15 AMPs. But I started speculating, and thought about:

  • "missiles", say bishop or rook type pieces that move about 10/turn [maybe unlimited slide - maybe! and may or may not have/need a leader.]
  • Multi-leader AMPs, with specialized organs within a larger transport/defensive shield simple AMP. These organs may:

grow missiles
contain/be a cluster of medium-range pieces that can "explode on contact" - see "Daisies" in the CVwiki.
maybe grow and drop "bombs"
exist independently

Enough for now. I'm going to put most of this email on the wiki

** Discussion 07/27/08 **

Distributed AMPs

Spread the units of each AMP out across the board. Drop the close proximity to move rules. Use either a great command control radius, like 10 squares, say, or drop any requirement for command control at all, and just let all the pieces of any one color move, wherever they are, as long as that color's leader is still on the board.


The current units in Chesimals 1 are guards, knights, modern elephants, modern warmachines, and bent sliding generals. There are currently 3 make-up for chesimals:
1 - the "ortho": 4 guards, 4 warmachines, 1 general
2 - the "dia": 4 guards, 4 elephants, 1 general
3 - the "royal": 4 guards, 2 knights, 1 royal general

A strange way to play Chess with Different Armies is to create some mixed chesimals. For example, you could have:
4 - the mixed1: 4 guards, 2 elephants, 2 warmachines, 1 leader.
5 - the royal supermix: 2 knights, 2 elephants, 2warmachines, 1 leader.

Update 4/14/09

Some random notes:

Extensive playtesting of the larger variants, Superchief and Overlord, have shown they do scale beautifully, and could get larger. Adding the 2 extra pieces to the original 5 in Overlord has shown that there is a bit of a learning curve both for going to larger sizes and more types of pieces, but with a game or two, Overlord is readily, even easily, playable. The next size should be in the 15x50 to 20x60 range, with 100 or so pieces. There need to be more piece types, also. We could even go whole-hog and throw in a new [higher-level?] leader.

Playtesting chesimals with Carlos Cetina has indicated they might need somewhat larger boards, and that the exact placement of a chesimal is apparently at least sometimes critical in the "smaller" chesimals games- shifting by 1 square makes the difference between a game where one chesimal could charge across the board with a devastating attack, and one where that same first turn all-out winning attack was blocked. This alone indicated more space was needed.

Played the first "game" of "Corps Chess" with Ji on Sat 4/11 during a visit. We each played a "loose" chesimal - range of 2 between pieces for orders to transmit as opposed to the current "tight" chesimals, which must be touching [range of 1 to transmit orders].

I'd write more but I gotta take care of the dog! ;-)


The Engine is a new type of leader/piece that allows the geometric translation of a group of pieces from one area on a board to another while maintaining all pieces in the same relative locations they started with. This translation across the board does not count as piece movement. Any pieces that had not moved before the translation still have the option to move after translation. Pieces being translated which moved into the translation area earlier in that turn may not move again that turn.

Engines have a radius of effect that is measured in squares from the engine. This measurement may be made in 2 simple ways. One type measures orthogonally or diagonally, moving a square area. The other measures the number of squares only orthogonally, and thus translates a diamond-shaped area.

Engines may be true leaders, that is, self-activating. Or engines may require a leader, perhaps a specific kind of leader, to activate them. The second choice leads to more strategic considerations, in the sense that some pieces cannot move at the destination, specifically at least the engine and the leader that activated it. This requires more planning and gives more opportunity to make a mistake.

Engines would be characterized by area translated and distance translated. This might be fixed but could conceivably be flexible, with smaller areas translated farther across the board.

Could engines be chained? That is, could a large engine translate its area, including a medium and small engine, then the medium engine translates its area, which includes the small engine, which then translates its area, where active leaders and pieces could prosecute their moves? The basic rules would seem to allow or even encourage it. The complexity allowed would be awesome! It almost requires a defensive leader…


A defensive leader is one which you may only use/activate during an opponent's turn to provide a point defense type of response to a [complex] action.

Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-Share Alike 2.5 License.