Code that trades our own capital.
Defam runs a fully self-funded algorithmic book on our trading platform. Every algorithmic strategy we deploy trades our own money — no external clients, no managed accounts, no signal subscriptions. The discipline below is what we use on our own desk.
An honest workhorse.
our trading platform is the platform we use to run live algorithmic strategies across FX majors, indices, metals and selected CFDs. It is broker-agnostic, fully scriptable in our scripting language, and connects cleanly to our external Python research stack via sockets and ZeroMQ.
For a self-trading firm, our platform sits at a pragmatic sweet spot: mature scripting language, deep broker coverage, low-latency execution against ECN venues, and the ability to host Expert Advisors on dedicated VPS infrastructure without vendor lock-in we cannot replicate in code.
We are deliberate about what our platform is not for. US equities and most listed futures live in different stacks. But for a globally diversified FX-and-CFD self-trading book, our platform remains, in 2026, the most operationally honest platform available.
How a live algorithmic stack is structured.
Python — pandas, vectorbt, a custom event-driven simulator for path-dependent logic, and notebooks for every hypothesis we test. Nothing reaches production without surviving walk-forward on out-of-sample data.
A portfolio of algorithmic strategies on our platform. Each strategy encodes a single, well-defined edge with hard-coded risk limits. No shared state between strategies — if one breaks, it breaks alone.
An external risk daemon monitors every account in real time. It can flatten positions, freeze new entries, or kill a strategy outright if exposure drifts outside tolerance. The trader cannot override it intraday.
The supervision layer is not a chart overlay — it is a separate service with its own audit trail. We describe how we build it on the risk management page.
See risk management →Strategy families.
Each family below maps to one or more algorithmic strategies on our live our platform stack. Signal generation is documented separately.
How we build signals →- Cross-asset momentumTerm-structure and trend signals across FX majors, indices, and metals.
- Statistical arbitrageMean-reversion pairs and baskets on correlated FX and index instruments.
- Volatility & carryCalibrated exposure to short-vol and carry premia in selected EM pairs.
- Event-conditional executionPre-defined behaviours around scheduled macro events — no improvisation.
- Liquidity-aware market makingProvisioning around news lulls in deep, well-spread instruments only.
Where this connects.
Want the technical detail?
Read the long-form article on building an our platform algorithmic stack.