Enter your capital and pick a risk tolerance — the engine builds a portfolio from top-ranked stocks using alpha-tilted inverse-volatility weighting (Treynor-Black 1973, the same logic Bridgewater uses in All Weather for $150B AUM). Per-position cap, per-sector cap, minimum position size — all institutional standard.
$
◆ Plain English
We balance the portfolio so volatile stocks get smaller positions and steadier ones get bigger. On top of that, we tilt slightly toward the names the engine ranks as the strongest buys.
What this means for you: a smoother ride than blindly putting equal dollars into every pick, with the same upside potential — because the engine's top picks still drive most of the return. Real-money use needs a 30-day paper-trade run first; see Virtual Portfolio.
◆ How the math works
Each candidate gets a raw weight proportional to its expected alpha divided by its volatility — that's the Treynor-Black ratio: more return per unit of risk = more room in the portfolio. Then the constraints kick in:
Per-position cap — no single ticker takes more than N% of the portfolio. Conservative 6%, Balanced 10%, Aggressive 18%.
Per-sector cap — no sector exceeds M% (protection against single-industry concentration).
Min position — positions below N% are dropped (transaction cost wouldn't pay off).
Min Framler score — bearish tickers are excluded automatically.
Volatility estimate — taken from the conformal interval width (the engine's own forecast uncertainty).
⚠ Not investment advice. The engine is a research tool. Every expected return is a prior-mode estimate until 2026-05-16; after that the engine flips to posterior-mode based on measured forward returns. The backtest is still warming up — see /backtest. Before any real trading, paper-trade through Virtual Portfolio for at least 30 days.