Lab

same town, different rules: how do they turn out?

Runs

Each line is one run (the live town, or a copy running with different settings), averaged across its raccoons, by sim date. A fork carries the history it was copied from, so its line sits on top of the original until the moment they split.

Settings per run

Start an experiment

Then on the Pi (snapshots so far: none):
cd ~/peanut-raccoon && source .venv/bin/activate
python -m peanut.lab fork before-storm harsh --set novelty_ratio=12
python -m peanut.lab fork before-storm gentle --set novelty_ratio=2
python -m peanut.lab run harsh --days 14 --fast --mock      # free, as fast as the Pi can go
python -m peanut.lab run gentle --days 14                  # real time with Claude (costs tokens)
Settings you can change: . A fork never touches the real town.