> METHODOLOGY

How the Orange Score is built

The public score is simple by design. The engine underneath it is not. This page explains the major layers without requiring you to read the full mathematical paper.

From many signals to one readable score

Orange Terminal's public interface deliberately starts with a simple result. Underneath that result, the engine builds a current market state, fits multiple current-conditioned future-path candidates, groups those paths into broad scenarios, and evaluates how the present state compares with historical conditions.

The main score components

Condition normality

How familiar or unusual the current state looks compared with the stock's own historical behaviour.

Trend quality

Whether recent movement is firm, weakening or inconsistent across time windows.

Expected volatility

Whether modeled price swings look controlled or stressed.

Price resilience

How well recent price action has held up, including drawdown context.

Scenarios and confidence

The engine does not rely on one deterministic future path. It evaluates multiple plausible paths and groups similar outcomes into scenarios. The amount of agreement among those scenarios becomes a separate confidence reading. This separation is intentional: a stock can have a reasonable health score while the future remains unusually uncertain.

Open methodology

Orange Terminal is built around open research and an open-source codebase. For the mathematical detail, read the white paper or inspect the GitHub repository.

CHECK A STOCK NOW

See the Orange Score in action.

Open a stock report to see its current 0–100 reading, plain-English state, scenarios and diagnostics.

CHECK SPY'S ORANGE SCORE