Memo is generated deterministically from the fund's ledger, audit log, benchmark snapshots, agent decision bundles, and scorecard evaluations. No LLM is involved in the prose; identical inputs always produce the identical memo. Risk-free rate assumed zero for Sharpe and Sortino — these figures overstate risk-adjusted return relative to calculations using the contemporaneous T-bill rate. Performance figures are not GIPS-compliant. "Excess Return vs benchmark" is the arithmetic spread between fund and benchmark total returns since inception, not a CAPM alpha. Theme contribution and contributor / detractor figures are unrealized mark-to-market — they will move with prices and may not match realized P&L when positions close. The closed-position win rate counts round-trips by ticker (FIFO matching) and is not size-weighted; a single large loss can outweigh many small wins. The agent scorecard's hit rate and mean score return are internal model artifacts measured against forward prices on the agent's stated horizon — they are not realized returns to the fund and should not be read as such. Trading signals are generated in part by LLM outputs, which may be incorrect or systematically biased; the operator assumes all responsibility for trade decisions. The fund holds concentrated positions in volatile instruments including cryptocurrency-linked securities. Universe, risk policy, and operational parameters are subject to change without notice. Nothing on this page constitutes investment advice or solicitation.