How Horizon works

Marketing/Content note: Initial scaffolding — refine the explanation and add a diagram of the idea → backtest → live flow.

Horizon has three stages, and you move through them in plain English.

1. Describe

You describe a strategy in natural language. Horizon's AI translates it into a precise, transparent set of rules — entries, exits, indicators, and risk controls. You can always see and refine the rules.

2. Backtest

The rules run against historical market data on an event-driven engine that respects the order of events (no look-ahead bias). You get standard performance metrics to evaluate the edge. See Backtesting.

3. Execute

When you're satisfied, the same rules deploy to live markets through a connected broker. Execution is automated and applies your risk guardrails.

Why plain English

Traditional algorithmic trading requires Python, Pine Script, or MQL. Horizon removes that barrier while keeping institutional-grade backtesting rigor — so the strategy you test is the strategy you run.