Abstract
Each cell of an infinite grid is alive or dead, and the next generation is decided by the eight neighbours alone: a live cell survives if $2$ or $3$ neighbours are alive, and a dead cell is born if exactly $3$ are (Conway’s Game of Life). Each of the twelve shapes made of five cells (the pentominoes) is placed alone, and the generations until it settles are counted. No new theorem or law is claimed.
First, the R-pentomino does not settle for $1103$ generations——the population begins to repeat with period $2$ at generation $1103$, with $116$ cells (agreeing with the known values). The largest population on the way is $319$ cells at generation $821$. From generation $1103$ to $1400$ the population stays $116$, but the rectangle enclosing the live cells keeps widening from $525$ by $501$ to $673$ by $649$ (Section 2).
Second, the other eleven settle within $10$ generations——counting the first generation at which the board itself equals the board two generations earlier, the longest, T, takes $10$. Five, N, P, U, Y and Z, die out; V and W become $7$-cell still lifes; and four, I, L, T and X, become $12$-cell oscillators of period $2$ (Section 3).
Third, judged by population, V and W come out one generation early——counting the generation at which the population begins to repeat with period $2$, nine shapes give the same generation as the shape criterion, and only V and W come out $1$ earlier. W has $7$ cells from generation $1$, but its shape changes until generation $2$. The population is a shadow of the shape, not the shape itself (Section 3).
Fourth, and this is the core. With the same five cells, shape splits the lifetimes into within $10$ generations and $1103$——the rules and the size are the same; only the arrangement of the cells differs (Section 4).
Keywords: Game of Life, R-pentomino, cellular automaton, glider, pentomino.
Scope (what is not claimed)
No new theorem or law is claimed——the rules of the Game of Life (Conway, introduced by Gardner in $1970$), that the R-pentomino settles to $116$ cells at generation $1103$, and that a glider moves one cell diagonally in $4$ generations are all known. Theories such as universal computation are not treated. Why the R-pentomino lasts so long is not asserted——only the generations, populations and spreads are shown. The settled shapes are not named——only their cell counts, and whether each is a still life or a period-$2$ oscillator, were checked. Relation to earlier papers: Paper 621 counted that under the $3n+1$ rule, $27$ takes $111$ steps and rises to $9232$ on the way——here, under a two-line rule, the five-cell R takes $1103$ generations and grows to $319$ cells on the way. What is added is setting all twelve side by side under the same rules, matching the shape criterion with the population criterion and finding the two shapes where they differ, and placing the separator on shape.
1. Notation and the Question
Rules: the neighbours of a cell are the eight cells around it, horizontally, vertically and diagonally. A live cell lives on if $2$ or $3$ neighbours are alive, and otherwise dies. A dead cell is born if exactly $3$ neighbours are alive.
Two criteria: the shape criterion takes as the lifetime the first generation at which the board equals the board two generations earlier; from then on it is a still life or a period-$2$ oscillator. The population criterion takes the first generation at which the number of live cells begins to repeat with period $2$ and continues for $100$ generations. If something flies away keeping its shape, the shape criterion cannot be used, but the population criterion can.
Question: how many generations do the twelve five-cell shapes take to settle?
2. The R-Pentomino
Computation 2.1: the R-pentomino (the F shape) was run for $1400$ generations. The population begins to repeat with period $2$ at generation $1103$, with $116$ cells. The largest population on the way is $319$ cells at generation $821$. Both agree with the known values.
| generation | height | width | population |
|---|---|---|---|
| $1103$ | $525$ | $501$ | $116$ |
| $1200$ | $573$ | $549$ | $116$ |
| $1400$ | $673$ | $649$ | $116$ |
Observation 2.2 after generation $1103$ the population stays $116$, but the rectangle enclosing the live cells keeps widening. Something moves away keeping its shape——so the shape criterion cannot be used for the R-pentomino, and the population criterion was used.
Computation 2.3: running the five-cell glider for $4$ generations returned the same shape shifted one cell diagonally. It is one example of something that moves keeping its shape.
3. The Other Eleven
Computation 3.1: each of the other eleven was placed alone, and its lifetime counted by both criteria.
| shape | shape criterion | population criterion | final cells | settled as |
|---|---|---|---|---|
| I | $6$ | $6$ | $12$ | period-$2$ oscillator |
| L | $9$ | $9$ | $12$ | period-$2$ oscillator |
| N | $5$ | $5$ | $0$ | died out |
| P | $4$ | $4$ | $0$ | died out |
| T | $10$ | $10$ | $12$ | period-$2$ oscillator |
| U | $4$ | $4$ | $0$ | died out |
| V | $3$ | $2$ | $7$ | still life |
| W | $2$ | $1$ | $7$ | still life |
| X | $6$ | $6$ | $12$ | period-$2$ oscillator |
| Y | $3$ | $3$ | $0$ | died out |
| Z | $3$ | $3$ | $0$ | died out |
The longest is T, at $10$ generations.
Observation 3.2 the population criterion comes out one generation earlier than the shape criterion for V and W. W has $7$ cells from generation $1$ onward, but its shape changes until generation $2$——a settled number does not mean a settled shape. The population is a shadow of the shape.
Observation 3.3 the two criteria give the same generation for nine shapes, and the final cell counts agree for all eleven. The shape criterion never comes out earlier than the population criterion——if the shape repeats every two generations, so does the count.
4. The Separator
Observation 4.1 this is the core: the separator is shape, not size. All twelve have five cells and follow the same rules. Eleven settle within $10$ generations, and only the R-pentomino takes $1103$——the only difference is how the five cells are arranged.
Observation 4.2 beside Paper 621. In Paper 621, under the $3n+1$ rule, $27$ took $111$ steps and rose to $9232$ on the way——here, under a two-line rule, the five-cell R takes $1103$ generations and grows to $319$ cells on the way. In both, a short rule does not promise a short journey.
Observation 4.3 what is not said. Why the R-pentomino lasts so long is not asserted, theories such as universal computation are not treated, and the settled shapes are not named——what is shown is the lifetimes of the twelve and where the two criteria agree and differ.
5. Roll Call and Closing
- R settles at generation $1103$ with $116$ cells (Computation 2.1): agrees with the known values.
- the largest is $319$ cells at generation $821$ (Computation 2.1): from five cells.
- the population stays while the spread grows (Observation 2.2): something moves away.
- a glider moves one cell diagonally in $4$ generations (Computation 2.3): keeping its shape.
- the other eleven within $10$ generations (Computation 3.1): the longest is T.
- five die out, two become still lifes, four oscillate (Computation 3.1): $0$, $7$ and $12$ cells.
- V and W come out one generation early by population (Observation 3.2): the count is a shadow of the shape.
- final counts agree for all eleven (Observation 3.3): two criteria.
- shape separates them (Observation 4.1): the core.
- a short rule, a long journey (Observation 4.2): Paper 621.
Closing: Under the two-line rules of the Game of Life, the five-cell R-pentomino does not settle until generation $1103$, grows to $319$ cells on the way, and leaves $116$ cells. With the same five cells, the other eleven settle within $10$ generations——five die out, two become $7$-cell still lifes and four become $12$-cell oscillators of period $2$. The separator is shape, not size. Judged by population, only V and W came out one generation early——a settled count does not mean a settled shape. Placed among the earlier papers——like the detour of $27$ in Paper 621, a short rule does not promise a short journey. To be honest——why the R-pentomino lasts so long is not entered, and the names of the settled shapes are not checked.
References
- Berlekamp, E. R., Conway, J. H., & Guy, R. K. (1982). Winning Ways for Your Mathematical Plays. Academic Press.
- Gardner, M. (1970). Mathematical games: The fantastic combinations of John Conway’s new solitaire game “life”. Scientific American.
- Yamagishi, Y. (2026). The Collatz Map Shrinks on Average, but Replacing 3 by 5 Makes It Grow [Paper 621]. Zenodo. 10.5281/zenodo.23014516
Reference
Yamagishi, Y. (2026). The Five-Cell R-Pentomino Does Not Settle for 1103 Generations, While the Other Eleven Settle Within Ten [Paper 628]. YOI: https://www.yamagishiyuuki.com/yoi/628
DOI: 10.5281/zenodo.23014546
ORCID: 0009-0006-8587-3450
PDF (English) · PDF (Japanese) · Date proof of the English PDF (OpenTimestamps .ots) · SHA256SUMS
The record page for this paper (in Japanese, with the PDFs in each language and the date proof): YOI 628. This paper is a preprint and has not been peer reviewed.