Mathematics Series · Cellular Automata · Est. 1970

Conway's
Game of Life

"Four rules. Infinite complexity. The universe in miniature."
4
Rules that generate everything
1970
Martin Gardner, Scientific American
Possible configurations
Turing-complete
Proven by Gosper's Glider Gun, 1970
The Four Rules — Applied Every Generation, Every Cell
Birth — A dead cell with exactly 3 live neighbors becomes alive.
Survival — A live cell with 2 or 3 live neighbors survives.
Death (isolation) — A live cell with fewer than 2 neighbors dies.
Death (overcrowding) — A live cell with more than 3 neighbors dies.

Play it yourself.

Click or drag to draw cells. Press Play and watch the rules take over. Load famous patterns from the library below, adjust the grid size, or go full-speed and see how quickly emergence kicks in from random initial conditions.

Conway's Game of Life — Interactive Grid
Speed: 10 gen/s
Grid:
Gen 0 · Pop 0

Can You Beat Bob?

Bob starts with 8 cells. You get 8 cells too. Place them anywhere on a 40×40 grid, hit Run, and the rules decide the rest. Beat Bob in longevity, peak population, or final stable population — or all three.

What's a Methuselah? In the Game of Life, a Methuselah is a small starting pattern that lives far longer than you'd expect — expanding, churning, and collapsing over hundreds of generations before settling into stable still lifes and oscillators. Bob's seed is one of them: 8 cells that peak at 179 cells and survive 338 generations before stabilizing. The name comes from the biblical patriarch who lived 969 years. Mathematicians have found Methuselahs with just 5 cells that run for over 1,000 generations (R-pentomino: 1103 gens). A comprehensive catalogue lives at LifeWiki → Methuselah.
Your Grid 40 × 40 · NO WRAPAROUND
Cells placed:
0 / 8
Rule: your live cells must fit inside a 5×5 box — any position on the grid is fine. Small constraint, surprisingly long lifetimes.
⚠ Your pattern must fit inside a 5×5 square — that's the rule.
Bob's Run PRECOMPUTED
Gen 0
Generations (stable) 338
Peak population 179
Final population 78
Link copied!

John Conway and the Machine That Thinks for Free

John Horton Conway spent two years in the late 1960s searching for the right rules. He wasn't trying to build a simulation — he was trying to settle a mathematical conjecture posed by John von Neumann: could a machine reproduce itself?

Von Neumann had already proven self-replication was possible in theory, but his model required 29 states per cell and thousands of cells to get started. Conway wanted something simpler. He tried hundreds of rule combinations, playing with boards of Go stones on his office floor, testing them by hand. He needed rules that were balanced — not too explosive, not too dead, complex enough to sustain long-term behavior.

When the rules finally clicked — Birth at 3, Survival at 2 or 3 — the game took on a life of its own. Conway brought it to Martin Gardner, who published it in Scientific American in October 1970. Within months it had consumed significant fractions of the world's computing time.

Why These Specific Rules?

Conway's intuition was about balance. Most rule sets collapse:


Too many birth conditions → explosive growth fills the grid and freezes. Every cell alive, nothing moves.


Too many death conditions → patterns collapse in a few generations. Nothing survives long enough to become complex.


Rules B3/S23 hit the sweet spot. Sparse enough that most patterns die, but rich enough that some structures find stable orbits. Life exists right at the edge between order and chaos — the same mathematical space where complexity always lives.

Conway tested thousands of rule combinations before arriving at B3/S23. The game bears his name, but the rules were discovered, not designed.
1948
Von Neumann poses the question: can a machine with simple local rules reproduce itself? Begins theoretical proof using 29-state cells.
1968–1970
Conway experiments at Cambridge, testing rule combinations on Go boards. Hundreds of candidates rejected. B3/S23 survives.
October 1970
Martin Gardner publishes "Mathematical Games: The fantastic combinations of John Conway's new solitaire game 'life'" in Scientific American. It becomes the most-read column Gardner ever wrote.
November 1970
Bill Gosper at MIT discovers the Glider Gun — a pattern that produces gliders indefinitely. Conway had offered $50 to anyone who could prove or disprove that any pattern could grow forever. Gosper wins it.
1982
Conway proves the Game of Life is Turing-complete. Any computation that can be performed by a computer can, in principle, be performed by patterns in the Game of Life.
2010
A working Game of Life computer is implemented inside the Game of Life — a grid running the rules of Life, with Life patterns forming its logic gates. Recursion complete.
April 2020
John Conway dies from COVID-19, aged 82. He leaves behind the surreal numbers, Monstrous Moonshine conjecture, free will theorem, and the most influential cellular automaton ever discovered.

The Bestiary of Life

In 50 years of study, thousands of patterns have been catalogued. They fall into families with precise names. Still lifes don't change. Oscillators repeat. Spaceships move. Guns produce spaceships. Each one is a mathematical object with known properties — period, bounding box, minimum population.

Gosper's Glider Gun and the Proof That Changed Everything

Conway believed most patterns in the Game of Life would eventually die or stabilize. He suspected no finite pattern could produce an unlimited number of cells — an "infinite growth" pattern. He offered $50 to anyone who could prove him wrong.

In November 1970, Bill Gosper and his team at MIT found it: a 36-cell pattern that fires a glider every 30 generations. Indefinitely. The Gosper Glider Gun.

This was not just a mathematical curiosity. Gliders can be used as signals. Two gliders colliding can be made to create or destroy structures. By designing careful collision patterns, you can build AND gates, OR gates, NOT gates — the complete set of logic primitives for a universal computer.

The Game of Life is Turing-complete. Every algorithm that can run on your laptop can, in principle, be computed by patterns of live and dead cells. The rules are a substrate for thought.

The Gosper Glider Gun — Live
Gosper Glider Gun — firing every 30 generations
Gen 0 · Gliders 0
Speed of Light in Life: Information in the Game of Life cannot propagate faster than one cell diagonally per generation — equivalent to c in physics. Gliders (the fastest natural "signals") travel at exactly c/4. This speed limit is fundamental, not a consequence of the rules — it's a geometric fact about the lattice. Every cellular automaton on a 2D grid has an equivalent speed limit.

Four Rules, Unbounded Complexity — the Ant Problem

The Game of Life and the Ant Supercolonies exhibit share the same mathematical heart. Both demonstrate that global intelligence requires no global design. Simple local rules, applied uniformly, generate behavior no one programmed and no one fully predicted.

Game of Life

Each cell looks at 8 neighbors and applies 4 rules. No cell has knowledge of the pattern it belongs to. No cell coordinates with distant cells. Yet gliders cross the grid, guns fire, self-replicators assemble. The "structure" exists only in the relationship between cells — not in any individual one.

Conway's universe is a proof: you don't need a designer for complex structures. You need the right rules.

Argentine Ant Supercolonies

Each ant responds to local pheromone gradients. No ant has a map. No ant knows about the colony spanning 6,000 km. Yet efficient food routes emerge, colony boundaries hold, and the supercolony coordinates millions of independent decisions into coherent behavior.

The colony is not in any ant. It's in the rules the ants follow — the same way a glider is not in any cell. It's in the rule B3/S23.

The deep question: Is the physical universe itself a cellular automaton? Edward Fredkin and Stephen Wolfram have argued it might be — that space, time, and matter are the emergent result of simple rules applied to a discrete substrate. Conway's game is the toy model for this hypothesis. Four rules. A grid. Ask yourself: how many rules does physics need?
→ See Ant Supercolonies Exhibit

Gardens of Eden and Self-Replicating Machines

Gardens of Eden

A Garden of Eden is a configuration with no predecessor — a pattern that could never arise from any previous state. It can only appear at the start of the universe.

John Moore proved in 1962 that Gardens of Eden must exist in any cellular automaton with the right properties. Roger Banks and colleagues found the first explicit Game of Life Garden of Eden in 1971 — a pattern so complex it required significant computer search to locate.

Gardens of Eden matter because they define the limits of what's "reachable" from natural starting conditions. They're the configurations outside the arrow of time.

Self-Replicating Structures

Von Neumann's original question — can a machine reproduce itself? — has been answered within the Game of Life. In 2010, Andrew Wade demonstrated a pattern called Gemini that destroys itself and reconstructs an offset copy.

The Gemini pattern contains a complete description of itself encoded in glider streams — exactly analogous to how DNA encodes the instructions for building the cellular machinery that reads DNA. The parallel is not metaphorical. It's the same mathematical structure.

Von Neumann needed 29 states per cell and decades of theory. Conway's 2-state system does it with simpler rules. Simplicity, again, generates more than we expect.

Life in the Wild — Where the Rules Show Up

The Game of Life is a model. But the same computational structures it demonstrates — emergence from local rules, self-organization without central control, computational universality in simple substrates — appear throughout the physical world.

🧬
Biology
Morphogenesis
Alan Turing's reaction-diffusion equations (1952) show how zebra stripes and leopard spots emerge from local chemical interactions — the same way Life patterns emerge from local cell rules. Biological pattern formation is cellular automata in continuous space.
❄️
Physics
Crystal Growth
Crystal formation follows rules applied locally at the growth front: temperature, pressure, bonding geometry. Snowflakes exhibit six-fold symmetry because water molecules follow the same rules regardless of where they're added. No blueprint required — only local information.
🧠
Neuroscience
Neural Firing Patterns
Early neural network models (McCulloch-Pitts, 1943) are formally equivalent to cellular automata. A neuron fires based on local inputs, applying a simple threshold rule. Consciousness — if it can be computed — emerges from these local interactions at massive scale.
🌊
Physics
Fluid Dynamics
Lattice-Boltzmann methods use cellular automata to simulate fluid flow. Each cell represents a local density and velocity; rules propagate these locally. The result: turbulence, vortices, and shockwaves — all from simple grid operations.
💻
Computer Science
Computation Theory
Wolfram's Rule 110 (a 1D cellular automaton with 3 states) is also Turing-complete — proved in 2004. Wolfram argues in "A New Kind of Science" that the entire physical universe may be computing itself through simple rules applied to a discrete substrate.
🐚
Biology
Mollusc Shell Patterns
The pigmentation patterns on cone snail shells can be reproduced exactly using cellular automaton rules — without any genetic instruction for the overall pattern. Each pigment cell reads neighbors, applies a rule, and deposits color. The pattern is emergent from the shell edge outward.

Conway Beyond Life

Conway reportedly disliked the Game of Life dominating his reputation. He had other work he considered more significant — and he was probably right. But the Game of Life is why most people know his name, so it earned its place in the history books.

Surreal Numbers

Conway invented surreal numbers during the analysis of Go endgames — a number system that contains all the reals, all the infinities, and all the infinitesimals, constructed from a single simple rule. Knuth described the discovery in a novella format in 1974. Mathematicians consider it one of the most beautiful algebraic structures ever found.

🌙

Monstrous Moonshine

Conway and Simon Norton discovered a mysterious connection between the Monster Group (the largest simple sporadic group) and modular functions in complex analysis — objects with no obvious relationship. The conjecture was proven by Richard Borcherds, who won the Fields Medal for it. The connection is still not fully understood.

🎲

Free Will Theorem

With Simon Kochen, Conway proved that if human experimenters have "free will" in the sense of making genuinely unpredictable choices, then elementary particles must also have this property. The theorem is a rigorous version of the intuition that quantum randomness is fundamental, not epistemic.

🔢

The Look-and-Say Sequence

Start with "1". Describe each digit: "one 1" → 11. Describe that: "two 1s" → 21. Continue: 1211, 111221, 312211... Conway proved this sequence always stabilizes into 92 fundamental "elements" that grow at a rate of Conways Constant: 1.303577... — a transcendental number hiding in a simple description game.

The Record

Primary Sources & Academic References
  • Gardner, M. (1970). "Mathematical Games: The fantastic combinations of John Conway's new solitaire game 'life'." Scientific American, 223(4), 120–123. — The original publication.
  • Berlekamp, E.R., Conway, J.H., & Guy, R.K. (1982). Winning Ways for your Mathematical Plays. Academic Press. — Contains Conway's formal analysis of Life patterns.
  • Wolfram, S. (2002). A New Kind of Science. Wolfram Media. — Chapter 2 covers cellular automata and computational universality.
  • Rendell, P. (2002). "Turing Universality of the Game of Life." In Adamatzky, A. (ed.), Collision-Based Computing. Springer. — Formal proof of Turing completeness via glider logic gates.
  • Banks, E.R. (1971). "Information Processing and Transmission in Cellular Automata." MIT Technical Report MAC TR-81. — First documented search for Gardens of Eden.
  • Turing, A.M. (1952). "The Chemical Basis of Morphogenesis." Philosophical Transactions of the Royal Society B, 237(641), 37–72. — Reaction-diffusion foundation for biological pattern formation.
  • Gosper, R.W. (1970). Unpublished correspondence with Conway. MIT AI Laboratory Memo. — Discovery of the Glider Gun.
  • Roberts, S. (2015). Genius at Play: The Curious Mind of John Horton Conway. Bloomsbury. — Definitive biography.
  • LifeWiki — Comprehensive database of Game of Life patterns, maintained by the community since 2009.
Share 🔗 𝕏