About

Conway's Game of Life is a zero-player cellular automaton devised by mathematician John Conway in 1970. The system operates on a two-dimensional grid where each cell exists in one of two states: alive (1) or dead (0). The next generation is computed by applying three deterministic rules based on the count of live neighbors N within the Moore neighborhood (8 adjacent cells). Miscounting neighbors or misapplying rules produces divergent evolution paths. A single misplaced cell can collapse a stable structure or prevent an oscillator from cycling. This tool implements the standard B3/S23 ruleset with toroidal boundary conditions, meaning the grid wraps around edges. It approximates an infinite plane within a finite buffer.

The simulation supports drawing cells directly onto the grid, inserting classic patterns, and controlling generation speed from 1 to 60 generations per second. Grid states persist across browser sessions via LocalStorage. Note: patterns near grid boundaries interact with the toroidal wrap. Structures that require unbounded space (e.g., infinite-growth patterns) will behave differently on a finite wrapped grid. Pro Tip: use the Step button to debug oscillator periods one generation at a time.

Formulas

Each cell C_x,y at position (x, y) transitions between generations according to its neighbor count N. The Moore neighborhood counts all 8 adjacent cells (horizontal, vertical, diagonal).

N_x,y = x+1∑i=x−1 y+1∑j=y−1 C_i,j − C_x,y

The transition function under the standard B3/S23 ruleset:

C_x,y^t+1 = {

1 if C_x,y^t = 0 and N = 3 (birth)1 if C_x,y^t = 1 and N ∈ {2, 3} (survival)0 otherwise (death)

Where C_x,y^t is the state of cell at position (x, y) at generation t, N is the count of alive neighbors in the Moore neighborhood, and boundary conditions are toroidal: indices wrap modulo grid dimensions W and H.

Reference Data

Pattern	Type	Period	Cells	Bounding Box	Discovered
Block	Still Life	1	4	2×2	-
Beehive	Still Life	1	6	4×3	-
Loaf	Still Life	1	7	4×4	-
Boat	Still Life	1	5	3×3	-
Tub	Still Life	1	5	3×3	-
Blinker	Oscillator	2	3	3×1	-
Toad	Oscillator	2	6	4×2	-
Beacon	Oscillator	2	6	4×4	-
Pulsar	Oscillator	3	48	13×13	1970
Pentadecathlon	Oscillator	15	12	10×3	1970
Glider	Spaceship	4	5	3×3	1970 (R. Gosper)
LWSS	Spaceship	4	9	5×4	1970
MWSS	Spaceship	4	11	6×5	1970
HWSS	Spaceship	4	13	7×5	1970
Gosper Glider Gun	Gun	30	36	36×9	1970 (B. Gosper)
R-pentomino	Methuselah	1103 (stabilizes)	5	3×3	1970
Diehard	Methuselah	130 (dies)	7	8×3	1971
Acorn	Methuselah	5206 (stabilizes)	7	7×3	1971 (C. Corderman)
Infinite Growth 1	Infinite Growth	∞	10	39×1	1971
Spaceship Flotilla	Composite	4	~30	Variable	Various

Frequently Asked Questions

This simulator uses toroidal boundary conditions. The top row wraps to the bottom row, and the left column wraps to the right column. A glider exiting the right edge re-enters from the left. This means edge-adjacent cells have neighbors on the opposite side of the grid, which can produce unexpected interactions if a pattern's debris reaches the boundary. On an infinite plane, these interactions would not occur.

B3/S23 means a dead cell is Born with exactly 3 neighbors, and a live cell Survives with 2 or 3 neighbors. Other Life-like automata use different birth/survival counts. For example, HighLife (B36/S23) adds birth at 6 neighbors, enabling a self-replicating pattern. Seeds (B2/S) has no survival rule at all, producing explosive chaotic growth. The notation Bx/Sy specifies the complete ruleset, where x and y are lists of neighbor counts.

The R-pentomino is a 5-cell pattern that takes 1103 generations to stabilize on an infinite grid, producing 116 live cells including 6 gliders. It was historically significant because it disproved the early conjecture that all small patterns either die or stabilize quickly. It demonstrated that simple initial conditions can produce complex, long-running behavior - a hallmark of emergent complexity in cellular automata.

Spaceship speed is measured as a fraction of c (the speed of light), where c equals 1 cell per generation - the maximum information propagation speed on the grid. A glider moves 1 cell diagonally every 4 generations, so its speed is c/4. The LWSS moves 1 cell horizontally every 2 generations at c/2. No finite pattern can exceed c due to the locality of the neighbor-counting rule. Constructing spaceships at arbitrary rational speeds below c is a deep open problem.

Yes. Conway's Game of Life is Turing-complete. Logic gates (AND, OR, NOT) can be constructed using glider collisions, and streams of gliders can encode binary data. Complete universal Turing machines and even pattern-based CPUs have been built within the Game of Life. This was proven by constructing universal constructors and was part of Conway's original motivation for studying the system.

Random soups with density near 37% (the critical threshold) tend to produce the most active evolution. After chaotic initial generations, most active regions collapse into stable ash composed of blocks, beehives, blinkers, and other small common objects. This is because these structures are robust local attractors. Statistical censuses of random soups show that blocks constitute roughly 50% of all ash objects, beehives about 25%, and blinkers about 10%. Larger or rarer structures (pulsars, gliders) appear with much lower probability.