Most strategies arrive as a mixture of a spreadsheet, a chart annotated in screenshots, and a set of rules the trader holds in their head. Our first job is to make all of it explicit, because the parts that stay implicit are exactly the parts that behave differently on a Thursday expiry with a gapped open.
What we build
Index and stock options strategies, futures systems, intraday equity models, event-driven books, portfolio rotation logic and multi-strategy runners that supervise several books at once. If you can describe the decision, we can encode it — including the discretionary overrides you want to keep.
How the code is built
- One source of truth. The same strategy class runs the backtest and the live session, so there is no second implementation to drift.
- Every decision logged. Including the no-trade days — a decision log that only records fills teaches you nothing about why the system stayed flat.
- Edge cases first. Partial fills, rejected orders, missing bars, stale quotes, post-close ticks, holiday calendars, changed lot sizes, symbol rollovers.
- Risk built in, not bolted on. Per-trade stop, daily loss cap, position and order-rate limits, flat-by-time rules, manual kill switch.
Deliverables
Source code in your repository with commit history, the signed specification, a test suite, a runbook, and a handover call. If the strategy will place orders through a broker API, we also supply the technical documentation your broker and exchange will ask for during algo registration.
We implement the client's own rules. We do not design trading advice, recommend instruments or strike selection as advice, and we quote no expected return for anything we build.