Strategy builder
Rules you can read,
months later
The value of a rule is that you can still tell what it does after you have stopped thinking about it. Conditions here stay legible: indicator, comparison, value, action.
- 3
- ways to compose the same strategy
- Readable
- every rule stays inspectable for the life of the strategy
Composition
Three ways in, one strategy out.
Describe what you want in words, start from a template and adjust it, or hand-edit each condition. All three produce the same object, so you can start one way and continue another.
There is no export step and no second representation. What you approve in the builder is what the backtester runs and what goes live.
- Plain conditions — indicator, comparison, value, action
- Nothing compiles into a shape you cannot inspect
- Switch between describing, templating and editing at any point
Before you deploy
Problems surface while you build.
Contradictions and unreachable rules are cheap to fix in the builder and expensive to find after a bad fill.
Found after deployment
- A condition that can never be true, discovered when nothing trades
- Two rules that contradict, resolved by whichever ran first
- A missing stop, noticed when the position is already open
Found in the builder
- Unreachable conditions flagged as you write them
- Contradictions surfaced before the strategy can be approved
- A required stop enforced by the rulebook, not by memory
Tested honestly
The same rules, run against history with costs charged.
Spread, commission and slippage are on by default. A backtest without them is marketing, not evidence.
Build one on a demo account.
Write a rule set, run it against history with costs charged, then let it trade a demo account and see whether it behaves the way it read.