GameOfLife Variations: From Classic Rules to New WorldsThe Game of Life, created by John Conway in 1970, is a cellular automaton that transforms simple local rules into unexpectedly rich and complex behaviour. Although the original ruleset—commonly called “Life”—has become a cornerstone of computational recreation and mathematical curiosity, dozens of variations, rule-sets, and worlds have been developed that change the dynamics profoundly. This article explores the classic rules, common modifications, notable rule-families, emergent phenomena, and how to experiment with and create your own Life-like universes.
1. The classic Game of Life: rules and essentials
Conway’s Game of Life is played on a two-dimensional grid of square cells that are either alive or dead. Time advances in discrete generations. The state of each cell in the next generation is determined by the current states of its eight neighbours (Moore neighbourhood):
- Birth: A dead cell with exactly three live neighbours becomes alive.
- Survival: A live cell with two or three live neighbours stays alive.
- Death: In all other cases, the cell dies or remains dead.
Despite the simplicity, Life produces a bewildering variety of behaviours: stable still lifes (blocks, beehives), oscillators (blinkers, pulsars), traveling patterns (gliders, lightweight spaceships), and engineered constructs like guns, breeders, and universal constructors.
2. Why vary the rules?
Rule variations let us explore different emergent behaviours, tune growth or decay, and discover new computational or artistic phenomena. Reasons to experiment include:
- Amplifying or suppressing growth
- Producing more chaotic or more ordered dynamics
- Creating different classes of mobile patterns or stable structures
- Investigating computational universality in alternative rule-sets
- Artistic or aesthetic exploration (textures, flows, colorizations)
3. Notation and families of Life-like rules
Life-like cellular automata are commonly described by the “B/S”notation, where B lists neighbour counts that cause birth and S lists counts that allow survival. Conway’s Life is B3/S23. Examples:
- HighLife: B36/S23 — same survival as Life, but a dead cell with 6 neighbours is also born.
- Seeds: B2/S — only birth on 2 neighbours; no survival rule, leading to explosive but simple patterns.
- Day & Night: B3678/S34678 — symmetric under cell complement; produces many stable, large-scale structures.
- Morley (aka Move): B368/S245 — supports interesting mobile objects.
- Life without death: B3/S012345678 — once born, cells never die; produces growing patterns and trees.
These changes can drastically alter long-term behaviour: some rules die out quickly, some explode, others support rich, stable ecosystems.
4. Key categories of behaviour
Different rules produce different macroscopic dynamics. A rough taxonomy:
- Class I — quickly settles to an empty or uniform state.
- Class II — forms stable or oscillating local structures.
- Class III — chaotic, pseudo-random patterns dominate.
- Class IV — complex, edge-of-chaos behaviour with mobile structures and possible computation (Conway’s Life sits here).
Which class a rule falls into often depends on births vs survivals balance, and whether births are possible in dense regions.
5. Famous variations and what makes them interesting
- HighLife (B36/S23): Adds the B6 birth. This small change allows a famous replicator pattern (a pattern that self-replicates), enabling self-reproduction experiments.
- Seeds (B2/S): Very explosive, creates fractal-like growth fronts. Useful for random pattern generation and visual texture.
- Day & Night (B3678/S34678): Has a “color” symmetry — flipping alive/dead yields the same rules — producing landscapes with large dark/bright regions and complex boundaries.
- Brian’s Brain (not strictly Life-like; 3-state CA): Cells have states: on, dying, off. Produces pulsing waves evocative of neural firing.
- Larger-than-Life: Generalizes neighbourhood radius beyond 1 (e.g., radius 2) with threshold rules; can model continuous-like patterns and large-scale “organisms”.
- Rule-space explorations (e.g., “LifeFinder” discoveries): Systematic searches through B/S space have found rules with guns, replicators, and universal computation analogues.
6. Emergent objects and their variants
Most Life-like worlds support analogues of:
- Still lifes: minimal stable patterns; their inventory differs by rule.
- Oscillators: periodic patterns — period lengths and shapes vary wildly.
- Spaceships: patterns that translate across the grid; different rules create different velocities and symmetries.
- Guns and puffers: sources of spaceships or trails; their existence often indicates richer computational possibilities.
- Replicators and breeders: patterns that copy themselves or produce growth — central to self-replication studies.
For example, HighLife’s replicator doubles every period and can combine to build programmable assemblers; Day & Night’s large stable regions allow “landscapes” with mobile boundaries.
7. Tools and environments for experimentation
- Golly: a powerful, cross-platform Life simulator supporting many rule-sets, scripting (Python, Lua), and large-scale simulation with Hashlife algorithm.
- Online applets and cellular automata sites: quick testing, sharing patterns.
- Custom code: implement CA rules in Python (numpy), JavaScript (for web visualizations), or shader languages (for GPU acceleration).
Example minimal Python (NumPy) step for a Life-like B/S rule:
import numpy as np from scipy.signal import convolve2d def step(grid, birth_set, survive_set): kernel = np.array([[1,1,1],[1,0,1],[1,1,1]]) neighbors = convolve2d(grid, kernel, mode='same', boundary='wrap') birth = (~grid) & np.isin(neighbors, birth_set) survive = grid & np.isin(neighbors, survive_set) return (birth | survive).astype(int) 8. Designing your own rule
Steps to create interesting variations:
- Choose a neighbourhood radius (1 for classic Life-like; larger for Larger-than-Life).
- Pick B and S sets. Small changes can have big effects—try toggling one neighbour count.
- Test random seeds, small crafted seeds (gliders, blocks), and patterns known from Life to compare behaviour.
- Observe classification: does it die, oscillate, explode, or produce mobile structures?
- Iterate; track patterns (still lifes, spaceships) you find and combine them.
Tips:
- If births occur at high neighbour counts, expect dense growth.
- If survival is rare, patterns tend to die quickly.
- Rules symmetric under complement (like Day & Night) produce striking landscape-like behaviours.
9. Visualization, color, and multi-state extensions
- Color mapping can encode age of a cell, velocity of moving objects, or different states.
- Multi-state automata (more than alive/dead) open neural, chemical, or ecological analogues: excitable media, wave propagation, predator-prey dynamics.
- Continuous-valued or probabilistic CAs blur discrete artifacts and model diffusion-like phenomena.
10. Computational and philosophical considerations
- Universality: Several Life-like rules are known or suspected to be Turing-complete because they support constructions analogous to logic gates, signal routing, and memory.
- Complexity from simplicity: Life variations illustrate how minimal local interactions can generate organized complexity— a metaphor used in biology, sociology, and complex systems science.
- Art and procedural content: Artists use CA rules for textures, patterns, and generative art; game designers use them for terrain and behavior simulation.
11. Example experiments to try
- Compare Conway’s Life (B3/S23) vs HighLife (B36/S23) from identical random seeds to see replicator-driven divergence.
- Run Day & Night from a noisy initial condition and watch large domains coarsen into stable lakes and islands.
- Create “seeds” experiments (B2/S) to grow fractal-like expansion and measure front speed.
- Search for small spaceships by brute-force enumeration or use communities’ pattern collections.
12. Communities and resources
Active communities maintain pattern databases, rule catalogs, and simulation code. Pattern collections (like LifeWiki) and forums are invaluable for discovering known objects and sharing new rule discoveries.
13. Closing thoughts
Exploring Game of Life variations is a doorway into emergent behavior, computation, and creative expression. A single modification to birth or survival counts can produce a new universe with its own physics, organisms, and aesthetics. Whether your goal is math, art, or playful tinkering, Life-like cellular automata reward curiosity with endless surprises.
Leave a Reply