STRATUM / METHODOLOGY
The rules behind the result.
Inspect the proposed research pipeline, the implemented scenario proxies and the boundaries of each model. Download the complete rulebook below.
01 / SYSTEM ARCHITECTURE
From market inputs
to defined exposures.
A five-layer research specification. Inspect the assumptions, transformations, and implementation boundaries at each stage.
Data qualification
A proposed eligibility layer rejects stale observations, incomplete contract histories and inconsistent timestamps before a signal is constructed.
- INPUT DOMAIN
- Reference prices · Contract terms · Funding curves
- OUTPUT CONTRACT
- Eligible universe + exclusion register
- PARAMETER REGISTER
- Draft gates: 252 observations; 20-session liquidity window; instrument-specific quote age.
Concept specifications describe a proposed process. Only the on-page deterministic calculations operate in this prototype.
Exposure & accounting map
Authored model assumptions| MODEL LAYER | REFERENCE EXPRESSION | RETURN CHANNEL | PRINCIPAL FAILURE MODE | IMPLEMENTED PROXY |
|---|---|---|---|---|
| Directional sleeve | Trend proxies | Price response | Momentum reversal / gap risk | 0.85 × spot shock + 0.10 × shock² |
| Volatility sleeve | Options-style carry proxy | Carry + volatility + convexity | Short-convexity loss / repricing | 7% × T − 0.004 × ΔIV − 0.70 × shock² |
| Basis sleeve | Cash–derivative spread proxy | Carry + spread movement | Funding mismatch / basis divergence | 5% × T − 0.35 × Δbasis / 10,000 |
| Financing ledger | Borrowing above 1.0× capital | Simple funding expense | Funding shock / collateral pressure | Capital × max(L−1, 0) × funding × T |
| Cost ledger | Management + one-time execution | Explicit deductions | Turnover / spread widening | Capital × fee × T + gross notional × cost |
02 / METHODOLOGY & CONTROLS
The assumptions
are part of
the product.
A research framework needs an explicit rulebook: what enters, what changes, how costs accrue, and where the model stops.
Download rulebookDefine the observation calendar, eligible instruments and data hierarchy before evaluation. A production design would distinguish exchange closes, composite marks and stale observations. The STRATUM demonstrator has no market-data feed: every path and coefficient is synthetic.