Dialog on a bike at night
The core of Mu's invention came on an eight miles nightly bike ride, from the game club Fanaat to my home, after having been exposed to Atlantis, a monumental but crooked game invented by Martin Medema. You can read more about the context in 'Organic Mechanisms'.

There are things about Mu that I sometimes have difficulty explaining:

  • In terms of structure it's a simple game.
  • I conceived it within an hour, without touching so much as a checker.
  • Its behaviour, though highly 'organic', is predictable enough to allow strategic planning.
  • Tactics may indeed backfire in over-the-board play. But the applet allows you to try out moves before submitting one.

The first pillar: the basic mechanism
The basic mechanism goes back to Sid Sackson's Focus and is based on a form of positive feedback: a column - a single included - moves as far as it is high. We're not concerned with bicolored columns, just with single-colored ones like these:

the basic organism

Here are some men arbitrarily divided over 6 columns. Consider it to be one organism. Like an ant colony it answers to a single mind: yours. It can split or merge, go this direction or that, crawl or jump, spread or erect, and it can display efficiency in that there's always a minimum number of step in which it can do things like:

  • Spread out completely.
  • Raise one stack consisting of all men.
  • Get at least one man to A.
  • Get say 10 men to B.
  • Occupy the area around C completely.

Or reach similar arbitrary objectives. It moves and morphs. For the moment it lacks growth ... but we'll get to that.

The second pillar: the basic terrain
The basic terrain goes back to an obscure seventies game in which a square would 'explode' if it would hold as many men as (or more men than) the number of its adjacent squares, ejecting one man to each of these, leaving the remainder behind, if any. It was called "Explosion" and was featured in issue 55 of Games & Puzzles Magazine. We had experimented with it before at Fanaat, including the hexversion. Martin's segments allowed for boards of different sizes and shapes and a 'one move per segment' protocol, that would enable move combinations without having them get out of hand.

Growth (will kill you)
Under the Atlantis explosion protocol, a column would 'explode' if its height reached the number of a cell's free neighbors, 6 on a centercell, 5 or less along the edge, limiting the columns' ranges accordingly. The game started with each player occupying one cornersegment filled with 7 men, one on each cell. It allowed each player to explode a first (capacity-3) cell on his second move.

Explosions provided the growing mechanism: If a cell exploded, it ejected one man to each of its neighbors, letting any remainder evaporate - that made me raise an eyebrow right away. Next the cell became a 'well', growing one man each turn untill it reaches capacity for a second time. Then it exploded once more and turned into a 'crater' - a solid obstacle. With wells and craters no longer counting as 'free neighbors', that, as it turned out, was a critical growth rate.

The negative side of positive feedback
Here's the thing about positive feedback: you have to keep it controlled or it will spin out of hand. Say you're a cell and your neighbor explodes: you get an extra man and at the same time lose a neighbor. Your capacity decreases while the load increases. That has "chain reaction" written all over it. Add that these chain reactions are most likely to creep inwards from the corners and edges, fueled by the wells, and the picture is clear: you're first and foremost trying to get away from your own wells, with whole sections along the edges eventually turning into craters and 'sinking into the sea' behind you ('cratered segments' were removed entirely). With any luck, you could secure some territory with targeted explosions, at a safe distance from one another, in the remains of what used to be a large board. I thought it was a rather pathetic object and a game that seemed designed with the sole purpose of hampering itself. The segments crumbling off the edges of the board still dominate my recollection of its first impression.
It was a strange night and I finally took my bike and went home in a state of confusion.

I must have been sleep biking on autopilot. I re-entered reality in the monochromatic orange light of an arterial road about half a mile from my home, about two o'clock in the morning. My legs were still peddling. Mu was born. I felt elated.

I had left Fanaat with conflicting impressions. On the one side there was this beautiful organism, versatile, fexible, efficient, volatile and capricious. What did it want? That was the key question. Certainly not the crippled fight to secure some space on a sinking island, fighting another sorry bunch of natives, driven onwards by explosions and crumbling edges of one's own making.
Not only the manner of erecting walls seemed wrong, but the place where they first appeared: in the players' own backyard, at the edge of the board. One needed growth, and low capacity edge cells were the only place to get it, initially. It was like building a wall against a wall. Meanwhile jumping to the center with high stacks to erect walls there, required making high stacks in the first place, without having them explode away accidentally in a chain reaction. In the center you'd need a 6-column for the first explosion, and next you'd need 5-columns for adjacent ones. But the attempted 'wall' would more often than not become an omni-directional 'blob' due to a chain reaction. Some way to build a wall.

Then, somewhere along the way, it occured to me that explosions are used, usually, to clear an area, not to erect something. And then the vision came. Ask any inventor how a game came to be, and it will probably be a more or less rational and evolutionary story along a timeline. But you can't rationalize a vision, or at least not its appearance. So I'll rationalize in retrospect, but it all happened in a fraction of a second.

A vision rationalized
To clear an area there must be something to clear in the first place. What I saw that instant was a board filled with a top-layer of white draughtsmen. Several separate 'holes' appeared in it and they grew bigger and bigger and inevitably encountered one another and ... didn't merge. Instead black draughtsmen appeared to replace any white ones the removal of which would cause a merger otherwise. An organically growing natural separation between different territories that, once completed, would leave a crude 'spiderweb' of boundary lines of black draughtsmen. Those of course would be the territories to conquer and defend.

It was all in one vision, one moment, and it included the growth of a new man on every cell that had its top-layer blown away. It provided fuel for the very same chain reactions, but with the reverse effect: they would actually clear one's territory instead of taking it away. A 'one-man-per-explosion' growth rate would be substantial, but not critical. This, I immediately felt, was what the organism was made for.

By the time I awoke, peddling, I had filled in most of the details: white draughtsmen would only have their own as neighbors, to determine a cell's capacity. The cells of territory they revealed when an explosion occured on them would have their own and those of the top-layer for neighbors. This would guarantee intricate interaction involving both layers. It would also allow any 'overcapacity' of an exploding cell to remain 'in place'. Much of the energy in Atlantis would evaporate as overcapacity, that had bothered me immediately. Of course cells that became part of the wall could also have overcapacity. I realized there would be men on the wall ... men that could travel the whole wall. Oh well, maybe not the whole wall, but they were there, and the question 'what would they want' had an obvious answer: remain involved. It was a detail that would solve itself, I felt. And it did, though it was not at all a 'detail', but rather the crucial key to invading territories! It goes to show once more that if the system is sound, the rule will be there.

The current protocol for laying out a board as part of the game, came later. It makes the game faster (due to a lower average capacity) with more room for opportunism based on local peculiarities. It also makes that 'the wall' does no longer necessarily consist of one connected web because it allows peninsulas to be locked off by seperate sections of the wall that have both sides terminated by the 'out of bounds'.

Mu velox © MindSports
Java applet © Ed van Zon