How AIBROKER Ranks Stocks

Step 1 — Universe construction

We use point-in-time index membership snapshots for the S&P 500, NASDAQ 100, FTSE 100, Nikkei 225, and EURO STOXX 50, plus curated baskets for sector ETFs, leveraged ETFs, inverse ETFs, single-stock ETFs, and crypto. A name is only eligible on a given date if it was a member of the index on that date — there is no survivorship bias.

Step 2 — Momentum scoring

Each eligible name receives a multi-horizon momentum score combining 1-month, 3-month, 6-month, and 12-month total returns, each volatility-adjusted. The score is then normalized within the universe to produce a percentile rank. The exact weighting and lookback windows are published and unchanged since launch.

Step 3 — Regime overlay

A two-state regime model (rolling realized volatility plus market-breadth) classifies every trading day as risk-on, neutral, or risk-off. The overlay does not change the per-name ranking — it tells subscribers when momentum has historically worked and when it has historically failed.

Step 4 — Cryptographic verification

Each daily ranking is concatenated, SHA-256 hashed, and the hash is published at 06:00 UTC before any US market opens. The full ranking is revealed after 21 trading days. Registered users (free account) can verify that the revealed ranking matches the original hash.

Backtest assumptions

• Point-in-time data only — no look-ahead, no survivorship bias.
• Daily rebalancing with realistic transaction costs (5 bps per side, plus a small-cap slippage adjustment).
• Equal-weighted top-K portfolios across the configurable holding horizons.
• All results are simulated. Past performance does not guarantee future returns.