A strategy and research company
We are building the research and operating foundation for a professional digital asset strategy platform.
Systematic digital asset strategies
Research first.Risk, always.
Zeyu Labs is a systematic digital asset strategy and quantitative research company. We develop rules-based portfolios around causal evidence, realistic execution and explicit risk limits.
Explore our approachResearch & institutional dialogue
01
Who we are
Institutional discipline for a market that never closes.
Our work sits at the intersection of quantitative research, portfolio management, risk control and trading engineering. The aim is not to forecast every market move, but to build decision systems that remain explainable under pressure.
01 / 02
This website does not claim licensed fund-manager status, external client assets, assets under management or a live client track record.
02
Investment approach
From a changing market to a controlled portfolio.
Each layer has a distinct job: define what was investable, rank the available evidence, size risk to the environment and constrain the final portfolio.
01Universe integrity
Point-in-time universe
The opportunity set is reconstructed as it existed at each decision date. Membership changes over time, reducing retrospective selection and survivorship bias.
02Asset evidence
Selection & trend
Cross-sectional evidence identifies a focused set of candidates. Trend confirmation helps avoid treating relative strength as sufficient on its own.
03Regime awareness
Dynamic exposure
A market-wide Gate adjusts risk progressively as participation weakens or recovers, allowing exposure to contract without relying on a single binary forecast.
04Risk allocation
Portfolio construction
Position limits, volatility-aware sizing, ATR exits and portfolio-level constraints turn individual signals into one governed allocation.
03
Research process
One chain of accountability.
A result only advances when the data, mechanism, implementation and failure modes can all be examined.
- 01→
Data
Timestamp what was observable and available.
- 02→
Hypothesis
State the mechanism before testing.
- 03→
Causal backtest
Separate signal time from executable time.
- 04→
Robustness
Challenge costs, windows and assumptions.
- 05→
Portfolio
Translate evidence into bounded exposure.
- 06→
Live controls
Fail closed when critical inputs are uncertain.
04
Validated evidence
Backtest / SimulatedFORMAL VALIDATION · 28 JUL 2026
Measured by return and the risk taken to earn it.
3,515.214%Total return
84.663%CAGR
−24.483%Max drawdown
1.801Sharpe ratio
3.458Calmar ratio
- Test period
- 22 Sep 2020 12:00 UTC — 28 Jul 2026 06:00 UTC
- Market data
- Binance spot public OHLCV · 6-hour bars
- Universe
- Semiannual point-in-time market-cap Top 10 · up to 6 holdings
- Execution
- Signals evaluated on completed bars · trades modeled at the next bar open
- Costs
- 0.10% taker fee + 0.05% all-in slippage per traded notional
Historical simulation, not live investment performance. Results include modeled fees and slippage but do not reproduce a full order book, queue position, market impact or every production constraint. Historical results do not guarantee future outcomes.
05
Risk discipline & technology
Risk is part of the model, not a paragraph at the end.
Portfolio rules and implementation controls are designed together so that research assumptions remain visible when a strategy moves closer to production.
Layered exposure
Market participation, asset-level trend and portfolio limits can each reduce risk. No single indicator has sole authority over the allocation.
Defined exits
ATR-based exits respond to asset-specific adverse moves, while position and portfolio constraints limit concentration before an exit is needed.
Execution realism
Signals use completed information, orders wait for the next executable window, and fees plus slippage are charged rather than treated as free rebalancing.
Shared rulebook
Research and controlled production interfaces are designed around the same core portfolio logic, with validation, safeguards and fail-closed behavior at their boundaries.
RESEARCH → PORTFOLIO → CONTROLS
A unified path, with deliberate separation of responsibilities.
Data availability, target construction, execution assumptions and operating safeguards are treated as named layers. This makes differences easier to audit without exposing private infrastructure or operational details.
Professional dialogue
Serious collaboration starts with a clear scope.
Zeyu Labs welcomes research exchange and conversations with quantitative researchers, market-infrastructure partners and long-horizon professional counterparties.
Professional enquiries · By established introduction