A one-person quant research desk that proves what doesn't work and ships the few things that do. — paper bots, ~108 tested-and-buried ideas, one real-money plan — all optimized for return per drawdown.
Flagships
All bots · click any to open its page
Backtests use each bot’s own history — spans vary from ~7¾ years (The Ark, 2019–26, incl. margin interest) to 16 years (Athena), and HOUSE v2.9 shows its 2015–26 after-tax backtest (the paper account itself started with $100k on Jan 2, 2026). Not directly comparable. The big number on each card puts every bot on one time frame: its rules replayed from $100,000 on Jan 2, 2026 to the latest close, before tax . The small line is each bot’s live paper account since its own launch.
30-year Monte Carlo · every bot · middle path from $100k · same method as each bot page
BOCA conviction · what we'd fund first, and why
Ranked by return-per-drawdown, robustness, and honest diversification — not raw return. The four voices don’t fully agree (Bill ranks Athena last; Athena ranks it near the top). See the full ranked reasoning + each voice’s dissent →
Explore the desk
The Map
Every bot, edge, and buried idea as a 3D force-directed graph — see how the whole desk connects.
Track Record
The desk's history and running results — what shipped, what died, and how the live book is doing.
Full Detail
Per-bot detail view with positions, trades, and the underlying account metrics.
The creed
Prove what doesn't work. Return is #1, drawdown is #2. Every strategy is tested survivorship-free, in- and out-of-sample, with no lookahead — most get buried, the survivors get deployed. The whole desk is watched by the HOUSE health bot (every 10 min), the Jasper security watchdog and a dead-man alert, and argued over by the BOCA council.
Not investment advice. Research and paper results only.