Does a Life soup evolve differently when growing regions get looser birth and survival rules? Green cells were born this tick and orange cells died this tick; dark cells have been alive longer.
Try this Lower cluster-threshold from 6 to 4 while it runs (it takes effect live) and compare with a default run. In the current build growth-clusters rises two- to threefold while live-cells does not change at all, because the cluster states never reach the cells.
About this model A MathPets experiment in multi-scale feedback: Conway's Life (B3/S23) plus two coarse-grained layers whose state feeds back into the cell rule. Every 4 × 4 block of cells is a cluster and every 12 × 12 block a district; growing clusters loosen birth, cooling clusters tighten survival, and blooming districts let crowded cells survive.
Each tick every cell counts its 8 live neighbors on the previous grid and applies B3/S23 with three changes: an empty cell with 2 live neighbors is born if its cluster is growing, a live cell with 2 dies if its cluster is cooling, and a live cell with 4 survives if its district is blooming. Then the 144 clusters and 16 districts are recomputed. A cluster is growing when it holds at least cluster-threshold live cells (default 6) or gained 3 since the last tick, and cooling when it holds half that or fewer, or lost 3. A district blooms when at least bloom-threshold (default 4) of its 9 clusters are growing. Both thresholds are read every tick; density applies at setup, shifted by +6 or −4 points in a checkerboard of districts.
Known issue: in the current build the zone lookup that copies cluster and district states onto each cell returns nothing, so cells follow plain B3/S23, the two thresholds do not change the board, and blooming-districts stays at 0. growth-clusters still counts clusters correctly.
