RISK
Drawdown-first
Smooth and durable over high and fragile. Capital preservation is objective number one.
We bring rigorous research, automated execution and a relentless focus on risk to the world's most liquid digital-asset markets — turning structural, well-documented market behaviour into a smooth, repeatable return stream.
We do not forecast prices. We identify persistent, structurally-grounded patterns in how markets price and transfer risk — and we capture them with rules, not opinions.
Edges in liquid markets are real, but fragile. They survive only for participants with the discipline to size them correctly, the technology to act on them without hesitation, and the humility to retire them the moment the evidence fades. Every position in our book exists because the data earned it a place there — never because of a narrative, a hunch, or a headline. Removing the human from the decision loop removes the single largest source of avoidable loss in active management.
Capital works across two complementary, low-correlation engines — one that earns by selling richly-priced risk, one that earns by following it. Together they aim to perform across regimes, not just in one.
Markets persistently overpay to insure against moves that mostly never come. We systematically sell that over-priced volatility — running a direction-neutral core whose maximum loss is defined and capped before entry. We collect the premium; the tail stays contained.
When markets move with conviction, we follow. Volatility-scaled stops let winners run and cut losers fast — the convex counterweight that pays for itself precisely when the rare, large move arrives and the income engine is tested.
Our entries and exits are gated by a blend of measures — some textbook, some proprietary. The components are no secret; which fire, at which thresholds, and in which combination is the part we keep behind the curtain.
Every measure is computed on live exchange data and validated out-of-sample before it is allowed to move a single position. Indicators that look good on paper but fail on real fills never make it into the book.
Nothing reaches live capital by accident. Every idea passes through the same disciplined funnel, and most are rejected along the way.
Hypotheses grounded in market microstructure and peer-reviewed literature — not folklore. We ask why an edge should exist before we ask whether it did.
Tested against real exchange data — live order-book and settlement prices — across multiple years and market regimes, with strict out-of-sample discipline to kill overfitting before it can cost a cent.
Every candidate is engineered with a known, bounded worst case before it is approved. If we cannot define the downside in advance, it does not trade.
Survivors trade live in simulation against real-time market data — proving themselves on tomorrow's data, not yesterday's — before touching a single dollar of capital.
Capital is allocated gradually and at a fixed fraction per position, with continuous reconciliation between model expectation and realised result.
Performance is watched around the clock. Edges that decay are cut — not defended. The research loop never closes.
We optimise for the shape of the equity curve, not the height of its peak. A strategy that compounds quietly beats one that spikes and collapses — and capital preservation is the precondition for everything else.
Positions are structured so the maximum loss is known and capped before entry. We carry no open-ended, undefined tail exposure.
Independent, low-correlation engines run side by side, so no single market regime can sink the book.
Each position risks a small, constant share of capital. No single trade — winning or losing — can dominate the outcome.
Conservative leverage, hard limits, and a willingness to forgo an outlier gain rather than risk a permanent loss.
Digital-asset markets trade every hour of every day. So does our infrastructure.
Live market data is ingested around the clock, and the system acts the instant a rule is triggered — no closed sessions, no overnight gaps, no missed moves.
Decisions are pre-computed and rule-bound; execution is mechanical, logged, and free of hesitation, fear or greed.
Every fill, every basis point of slippage, every deviation between model and market is recorded — and fed straight back into the research loop.
A modular research-to-execution stack lets us test, retire and add return streams without disrupting the live book.
The behaviours we harvest are well documented in the literature. Converting them into a smooth, durable return stream is a problem of process, technology and nerve — and that gap, between knowing and executing, is where our edge lives.