We build our own C++ engine, research repeatable edges over years of tick and order-book data, and deploy them across global sessions — on one codebase, where the research and the live engine are the same code.
Zed Quant Ltd · built and operated in-house for proprietary trading.
| waiting | Session-clock · index |
| in position | Overnight · Asia |
| waiting | Cross-market pair |
| done | Volatility burst |
| Session-close flatten | enforced |
| Per-strategy hard stop | armed |
| Max hold enforced | engine-level |
| Net exposure | 0 |
The operator's console — live strategy state and structural risk guardrails. Illustrative.
Zed X is a single low-latency C++ engine. The same code that researches a signal is the code that trades it — no reimplementation, no “it worked in the backtest” surprise. What we prove is what we run.
Backtest and live execution share one path. Given the same market data, the engine makes the same decision — every time. That parity is the point: research we can actually trust in production.
Research runs on years of tick and order-book data, GPU-accelerated, with fills modelled at the touch. Every candidate clears out-of-sample, real trading cost, and a multiple-testing null before it trades a single lot.
| Out-of-sample | ✓ |
| Per-year consistency | ✓ |
| Fills crossed at BBO + cost | ✓ |
| Multiple-testing / permutation null | ✓ |
| Session & roll rules applied | ✓ |
Every gate must pass before deployment. No single good year earns a lot.
Most backtests are noise wearing a good haircut. The discipline is structural: out-of-sample, per-year, and a date-block permutation null. If an edge is really just two lucky years, the pipeline says so — and it doesn’t trade.
Instrument × session × parameter combinations across multi-year tick data on GPU — hypotheses tested in minutes, not weeks.
Fills cross the spread at the BBO. Commission, slippage and exchange session rules are priced in before a strategy is ever believed.
A deliberately high bar for promotion. An edge has to clear the tests and still pay for itself after costs before it reaches live trading.
Trading futures around the globe and around the day — North America, Europe, Asia, and China’s onshore markets — anchoring signals to the exchange clock and the session that actually sets each market’s tone. Tokyo opens, London hands over, Chicago runs the afternoon.
Holding no risk is a deliberate choice, not an absence of one. Capital sits in cash unless a proven edge is firing.
Session-close flattening is enforced at the engine level, in backtest and live alike. Anything carried overnight is there by design, not by drift.
Every strategy carries a hard stop and a max hold, armed at the engine level. Guardrails aren’t something to remember — they’re built in.
Proprietary systematic trading and quantitative research, built and operated in-house. We write the C++ engine, conduct the research, and operate the live trading infrastructure.
The firm is focused on developing, deploying and operating systematic strategies across global futures markets.
For data, infrastructure and hosting conversations, and systematic-trading partnerships. Send a note and we’ll come back to you.