An experiment asking a real fly connectome to choose a Japanese equity allocation. Watch the decision, inspect the portfolio, and follow the results.
Loading the fund record
We simulate simplified firing dynamics on 176,422 neurons and 6,287,749 retained directed edges derived from MaleCNS. This is not a faithful biological brain emulation. The artificial readout learns; the fly connectome itself does not.
No brokerage trades are placed. This experiment does not establish profitability or predictive ability and is not investment advice. Monthly rebalancing inside each strategy is separate from the fly's allocation decision.
Annualized volatility and Sharpe ratio are omitted because the engine's observation frequency has not been validated.
Data: MaleCNS / FlyEM contributors (CC BY 4.0), via blendi-remade/fly-brain-minecraft. Modifications: connection normalization, simplified dynamics and an artificial readout. GitHub ↗