03 / Simulation

演算 12 个月。
不是拍脑袋。

Agent-based monthly simulation. 280 行 Python, 0 external dep. 你 clone 即跑。

BEAR
缓慢增长 / 用户流失 / 高 dispute
4,170
PACT GMV (12 mo)
最终 advertisers7
最终 creators53
总 campaigns41
cum burn58.5
cum council35
creator 总收3,783
BASE
7% MoM growth · 5% dispute
220,081
PACT GMV (12 mo)
最终 advertisers228
最终 creators2,505
总 campaigns2,182
cum burn3,235
cum council5,908
creator 总收209,203
BULL
40% MoM growth · 3% dispute
20.2M
PACT GMV (12 mo)
最终 advertisers7,820
最终 creators71,946
总 campaigns101,107
cum burn300,338
cum council378,661
creator 总收19.4M
CUMULATIVE · BASE SCENARIO

12 个月轨迹。
每月累积。

CREATOR EARNINGS (累积)
209,203
COUNCIL FUND (累积)
5,908
BURN (累积)
3,235
M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 Creator earnings
REPRODUCIBLE

多次演算。

改参数, 跑 seed, 对比 scenarios, 都是一行命令。

3 个 scenarios 全跑 + 对比
$ python3 scripts/economic_simulator.py --compare
N seeds (variance 估计)
$ for s in 1 2 3 42 100 999; do
    python3 scripts/economic_simulator.py --scenario base --seed $s
  done
5 年预测
$ python3 scripts/economic_simulator.py --scenario base --months 60
自定义 scenario (改 SCENARIOS 字典)
SCENARIOS["my_scenario"] = {
    "months": 24,
    "advertisers_init": 100,
    "advertiser_growth_mo": 0.15,
    ...
}
SIMULATOR ROADMAP

v0.2 / v0.3。

v0.1 · LIVE
Agent-based, 0-dep
· 3 scenarios CLI
· CSV 月度输出
· Bear / Base / Bull
· stdlib only
v0.2 · W5
Monte Carlo + price
· 1000 seeds P5/P50/P95
· PACT/USD oracle mock
· Competitor model
· Governance feedback loop
v0.3 · 长期
Live + interactive
· Jupyter notebook
· Grafana panel
· Calibration from chain data
· Web UI