Spaceships: How Five Cells Learned to Travel
Consider what it means for something to move in the Game of Life.
There is no physics here. No momentum, no velocity, no conservation of energy. Each cell lives or dies purely on the basis of its neighbors, one tick at a time. Motion, in the ordinary sense, is impossible: a cell cannot move from one square to another. It can only die in one place and be born in another.
And yet things move.
Not the cells — but the patterns. A pattern of live cells can, through a sequence of births and deaths, effectively translate itself across the grid. The cells that are alive at tick 4 are completely different cells from those alive at tick 0 — but the pattern they form is identical, just shifted in space. It has, in every meaningful sense, moved.
These patterns are called spaceships. They are among the most remarkable objects in the Game of Life, and one of them — the glider — changed the history of computing.
The Glider
The glider was discovered in 1969 or early 1970 by Richard Guy, a British mathematician who was corresponding with Conway about his new cellular automaton rules. It is a configuration of five cells that evolves through four distinct phases, returning to its original shape after four generations — but shifted one cell diagonally.
Generation 0 Generation 1 Generation 2 Generation 3 Generation 4
. # . . . . . . . . . . . # .
. . # # . # . . # # . . . . #
# # # . # # # # . . # # # # #(shifted one cell right and one cell down)
The glider is the smallest possible spaceship in Conway’s Life. It is also the most important.
Its importance is not aesthetic — though it is undeniably elegant. Its importance is computational. A stream of gliders, evenly spaced, can carry a binary signal across the grid. Two streams of gliders, aimed to collide, can interact in ways that implement logical operations. A glider that reaches a target pattern and destroys it is a signal that has done something. From these facts, it follows that gliders can be used to build computers — and eventually, the proof was completed that Life is Turing complete.
The glider did not cause this. But it made it possible to imagine it.
Why Spaceships Exist at All
The existence of spaceships is, in retrospect, not obvious. In fact, it is slightly surprising.
Consider the alternatives. A pattern could:
- Die out immediately (too sparse to sustain itself)
- Grow chaotically and without bound
- Stabilize as a still life (unchanging)
- Oscillate in place (an oscillator)
- Translate across the grid (a spaceship)
The first three happen commonly. Oscillation is somewhat less common but well-established. But translation — actual movement across the grid — requires a very specific kind of self-repair: the pattern must reconstruct itself just ahead of where it currently is, while simultaneously dismantling itself behind.
This requires the pattern to be, in a sense, asymmetric in the right way. Oscillators are typically symmetric. Spaceships are not. The glider’s five cells are arranged in a shape that is slightly lopsided — and that lopsidedness is what drives it forward.
Speed Limits
Spaceships in the Game of Life are subject to a fundamental speed limit.
Information in Life can travel no faster than one cell per generation — that is, “the speed of light,” by analogy with physics, written as c. A pattern cannot affect cells more than N cells away in N generations, because the causal influence of any cell propagates at most one cell per tick.
This imposes a hard upper bound on spaceship speeds. A spaceship moving diagonally can travel at most c/4 (one cell diagonally per generation; since diagonal movement covers $\sqrt{2}$ cells per generation of Euclidean distance, the convention is $c/4$ in the standard notation where movement of one cell per generation orthogonally = $c/2$). A spaceship moving orthogonally (horizontally or vertically) can move at most c/2.
| Speed | Direction | Description |
|---|---|---|
| c/4 | Diagonal | Maximum diagonal speed; glider is c/4 |
| c/2 | Orthogonal | Maximum orthogonal speed |
| c/3, c/4, c/5… | Orthogonal | Slower orthogonal ships |
For decades after the glider was discovered, finding spaceships of various speeds was an active area of research. Smaller, faster ships are harder to find because the space of possible patterns is vast and the constraints are tight.
The Standard Fleet: LWSS, MWSS, HWSS
Beyond the glider, three other spaceships were discovered very early and are now considered canonical:
LWSS (Lightweight Spaceship): 9 cells, period 4, travels orthogonally (horizontally or vertically) at speed c/2. It is the smallest orthogonal spaceship.
MWSS (Middleweight Spaceship): 11 cells, period 4, travels orthogonally at c/2. Similar to LWSS but slightly larger.
HWSS (Heavyweight Spaceship): 13 cells, period 4, travels orthogonally at c/2.
These three ships form the “standard fleet” — the basic tools with which Life engineers build larger systems. They interact in well-studied ways, can be produced by known guns, and serve as signals in computational constructions.
Glider Guns: Making Spaceships on Demand
For about six months after the Game of Life was published, no one knew whether it was possible to produce an infinite number of spaceships. Could you build a pattern that kept generating gliders, endlessly?
Conway offered a $50 prize for proof one way or the other. In November 1970, Bill Gosper’s team at MIT found the answer.
The Gosper Glider Gun is a period-30 oscillator — a pattern that returns to its original state every 30 generations — that emits one glider per period. It was the first configuration proven to enable unbounded growth in Life, and it settled Conway’s prize immediately.
The glider gun is important for the same reason gliders are important: it provides a source of computational signals. In a Life-based computer, you need a way to generate input — the glider gun provides it. Read more about guns →
Spaceships as Computation
The jump from “interesting moving pattern” to “component of a universal computer” is shorter than it might seem.
Consider what a computer does: it takes signals, routes them, processes them according to logic, and produces outputs. In Life:
- A stream of gliders is a binary signal (the timing of gliders encodes information)
- Glider streams can be created (using guns), destroyed (using eater patterns), and reflected
- Two glider streams can interact to produce AND, OR, NOT logic via specific collision configurations
- A “delay line” can be built to store signals by routing them through a long path of reflectors
With these components, you can build registers, counters, and eventually a full Turing machine — all from patterns of cells following four rules. Read the full proof →
The Broader Family
Outside of Conway’s Life, other cellular automaton rules produce their own spaceships — often with wildly different characters.
In HighLife (B36/S23), spaceships are rarer, but the rule is famous for its replicator — a pattern that creates copies of itself, including copies capable of further replication. In Day & Night (B3678/S34678), the symmetry between live and dead states means that the “negative” of any pattern is also a valid pattern, leading to unusual spaceship families. In Seeds (B2/S), there is no survival at all — every live cell immediately dies — so movement takes on an explosive, propagating character very different from the glider.
The general study of spaceships across rule-space — what speeds are possible, what shapes occur, what speeds require what minimum cell counts — is still an active research area. Explore the variants →
What Spaceships Tell Us
There is something philosophically interesting about the existence of spaceships.
The Game of Life has no physics. It was not designed to support movement. Conway was looking for an interesting rule, not a physics engine. And yet movement emerged — not as a feature, but as an inevitable consequence of the right combination of local rules.
This is, in miniature, a model of how movement might have emerged in the physical universe. In physics, particles are not “moving things” in a pre-existing space — they are patterns of field excitation that propagate through time. A photon is not a thing that moves; it is a pattern of electromagnetic oscillation that, by the geometry of Maxwell’s equations, necessarily propagates at c.
The glider is the Game of Life’s photon. It does not move because anything is pushing it. It moves because the rules, applied locally, inevitably produce something that looks exactly like movement.
Four rules. Five cells. One universe that turned out to have, among its natural phenomena, the concept of motion.