Methodology
The construction is protected. The proof is public.
Deterministic where it provably publishes.
Every score, every verdict, every projection from the Sovereign Risk Infrastructure is reproducible, auditable, and anchored to validation evidence – not to a black box. The mathematical method behind the ARIN22 kernel is published at the level of its mathematical class; the construction itself is a protected trade secret, disclosed only under agreement.
Proof discipline
Evidence before formula.
The publishable claim is always anchored to evidence, never to a secret construction. Public pages show the envelope; NDA packs show run evidence, hashes and replay bundles; signed engagements handle protected derivation.
Verified truth.
Every output traces to source data, snapshot, kernel version, routing class and replay hash. The result is inspectable by a model-risk reviewer.
Lane-qualified metrics.
Latency, parity and high-dimensional accuracy are always tied to a lane. D = 4 latency is not sold as D = 200 latency.
No silent degradation.
Hard regimes route to EVT / POT overlay or MC challenger lanes. The route is recorded before publication and carried into the audit trail.
Stochastic to deterministic
From path enumeration to deterministic tail metrics.
Classical Monte Carlo enumerates the future by brute force. ARIN22 reduces dependence on repeated enumeration inside bounded lanes and reproduces Monte Carlo tail metrics within published validation bands.
CVaR 99.9 within 0.035% of a 10M-path MC reference on the canonical lane; D = 200 high-dimensional lane max 0.90%. Deterministic where it provably publishes. EVT / MC challenger on hard regimes. No silent degradation.
The kernel method
What the kernel computes – and what stays sealed.
The public method describes the admissible class and the evidence envelope. It does not disclose protected construction, calibration thresholds or implementation details.
Named mathematical basis. Protected deterministic risk-state construction with EVT / POT tail overlay. The class is public. The construction is not.
Published authorship. The deterministic risk-compute method is published under the founder’s name, distributed across SSRN eJournals:
Anti-hallucination
Math governs the agents.
Risk numbers are computed by the deterministic kernel, never by an LLM. Agents explain, challenge and route. The Governor decides what can leave the system.
Deterministic constraint.
VaR, CVaR, stress and cascade numbers are reproducible from snapshot plus kernel version. Narrative cannot override the computed state.
Release-managed learning.
Realised-outcome recalibration enters through versioned calibration releases with replay tests and approval gates, never through silent runtime drift.
Single exit.
Before publication, a math-to-narrative audit blocks contradictions. The Governor is the single exit, and every verdict carries a cryptographic trail.
Latency to autonomy
Autonomy in regulated finance is a latency problem.
A council can deliberate and govern, but if risk revaluation takes minutes, no committee will trust the loop to act. ARIN22 validated lanes reduce core revaluation into sub-millisecond bands and move toward the full-stack 400 ms design target (Lane 3) – the measured value is published at contract signing, not before.
Execution remains client-side. SAA is not a broker, venue or execution engine. Precise benchmark configuration, speedup multipliers and compute-economics models are released to qualified partners under NDA.
Disclosure boundary
Public envelope. NDA evidence. Protected construction.
The page is intentionally explicit about what is public and what is not. Serious diligence gets evidence. Public readers get the claim boundary.
Public surface.
Latency envelope, accuracy bands, validation counts, class routing, regulatory mapping, integration surface and reproducibility contract.
Trade secret.
Protected construction, construction dossier, fallback logic, parameter calibration and kernel implementation.
TIER 1 – public page · TIER 2 – NDA pack · TIER 3 – signed engagement · No third-party validation attestation is claimed until issued.
Digital Assets
The SAA Institutional Rating Framework for digital assets.
Digital Assets applies the same discipline as the kernel: every number is computed by the deterministic stack from recorded inputs, every result carries its version, and a result without sufficient evidence is withheld rather than estimated. This section describes how assets are rated, how the SAA Qualified Universe is formed, and what the risk, portfolio and pre-trade figures mean.
Class-specific, versioned frameworks
An asset is rated under the framework of its class, not under a universal formula. Each framework fixes its pillars, their nominal weights (summing to 100), which pillars are load-bearing, a coverage floor and its grade thresholds. Every result names its framework and version (for example NATIVE_L1_POS_v2@v2), together with the scoring-model and bridge-policy versions.
Current frameworks for layer-1 networks (nominal weights, version v2):
| Pillar | Proof-of-Work L1 | Proof-of-Stake L1 | Other L1 |
|---|---|---|---|
| Fundamentals | 30 | 28 | 28 |
| Network Activity | 20 | 22 | 20 |
| Market Quality · Tradability | 20 | 20 | 22 |
| Security & Resilience | 12 | 17 | 15 |
| Regulatory Scope | 8 | 8 | 10 |
| Environmental | 10 | 5 | 5 |
Separate frameworks exist for DeFi protocols, infrastructure/middleware, meme assets and stablecoins. They use their own pillars (for example peg integrity and reserves for stablecoins, holder concentration for meme assets), and their grade thresholds are published only once calibrated. Layer-2 networks are not rated until their framework is ratified: they show "not computed by policy" rather than a provisional score.
What each pillar measures
- Fundamentals: network and token economics (fees, issuance or staking economics, security budget, utilization). Not a valuation verdict.
- Network Activity: observed on-chain usage from the network’s own data.
- Market Quality · Tradability: how tradable the asset is (order-book depth, spreads, turnover), not how much its price moved.
- Security & Resilience: consensus decentralization, the curated incident record with primary evidence, bug-bounty and client diversity. An absent record is "not assessed", never "clean".
- Regulatory Scope: the asset’s status in load-bearing jurisdictions (US and EU) from primary sources. If either is unassessed, the pillar is withheld. There is no favourable default.
- Environmental: energy and environmental context where material for the class. Heuristic and class-based; disclosed as such.
Evidence rules: withheld is a result
- Load-bearing pillars: if one is unmeasured, the composite score is withheld. It carries no number and no grade.
- Coverage floor: a composite requires at least 75% of its nominal weight to be scored. A missing supporting pillar is disclosed and the remaining weights are renormalized; it is never filled with an assumed value.
- Missing data cannot improve a score. Proxies, stale data and non-binding sources are labelled as such.
- Numbers come from the computational stack only. Language models write explanatory text and never produce or alter a figure.
SAA Qualified Universe
The assets available for analysis form the SAA Qualified Universe, a versioned list approved by SAA. Membership is decided by analytical coverage, not by market capitalization.
- Identity. The asset’s identity, network and data mappings are verified. Identity is necessary but not sufficient.
- Canonical completeness. The production rating is not withheld, and at least 75% of the framework’s nominal weight is scored.
- Analytical admissibility. Load-bearing pillars rest on admissible evidence: not proxy-only, not stale, not from non-binding sources.
- ACOS ranking. The Analytical Coverage & Observability Score orders qualified assets. Weights:
| Component | Weight |
|---|---|
| Model & data coverage | 30 |
| Market observability | 20 |
| Network / protocol observability | 15 |
| Evidence quality | 15 |
| Security & regulatory depth | 10 |
| Institutional analytical value | 10 |
| Market capitalization | 0 |
Changes happen only through an approved universe version after a qualification review. An asset enters after a passed review. A member leaves after two consecutive failed reviews, or immediately on a critical, identity or security finding. Individual results for a member can still be withheld when the evidence for that result is insufficient. Assets whose identity is verified but which are not yet approved are shown as "qualification in progress".
Risk figures
- Historical 1-day Value-at-Risk and Expected Shortfall come from the stated return history. The number of tail observations is always shown. Below five observations a figure is marked indicative. Deep-tail (EVT) estimates are published only when enough exceedances exist.
- Risk State reads current market conditions only: 30-day realized volatility, drawdown from the prior peak and the measured slippage of a $10M order, with open interest and funding as supporting drivers. It does not use the quality rating. A tail-validation result is reported beside it as model health, not as a risk level.
- Every figure names its sample: window, number of observations, data source, and the time of the last accepted data point.
Portfolio and pre-trade simulation
- Who computes: portfolio risk, dependence and scenario figures are computed by the SAA Risk Analyzer, a separately versioned compute authority. Digital Assets presents its verified receipts and computes none of these numbers itself. Results are released for review use.
- Construction: portfolios are constant-weight and rebalanced daily on historical data. Results are not realized P&L and not a forecast.
- Price panels: analysis runs only on price histories natively quoted in USD. If a native USD history is unavailable for an asset, the analysis is reported as unavailable. The asset is never silently dropped or substituted.
- Simulations only: pre-trade buy and sale simulations walk a single venue’s captured order book with a declared venue fee schedule. Holdings are inferred from weights and NAV, not read from an account. Sale proceeds are shown in USDT, a stablecoin exposure rather than USD cash. No order is ever sent, and nothing in the simulation constitutes best execution.
Data and source rights
- Market data: CoinGecko (attributed in the product) and venue order books.
- Network data: public chain endpoints and primary records for security and regulatory evidence.
- Source status: each source carries a recorded rights status (binding or transitional), disclosed with the result.
- Unavailable or insufficient data: the affected result is withheld, never replaced by a value from another source.
Disclosure boundary
Digital Assets provides descriptive analytics for research and risk review. It is not investment advice, does not issue trading signals or recommendations, and does not execute or route orders. No third-party validation or regulatory endorsement is claimed.
External evidence
External evidence
Published replay lanes, public-source arithmetic and withheld boundaries in one diligence ledger.
External evidence means we replay published figures from public sources through our kernel and publish the receipts. No regulator, index provider, or data vendor has reviewed or endorsed SAA. Our own loss models are withheld from these exercises by design.
| Lane | Public source | Result | Status |
|---|---|---|---|
| Changepoint detection (TCPD) | Turing Change Point Dataset, 15 methods | 4/15 mean-score (CI95 1-7) · earlier successful-rows frame 6/15 lies inside the CI – frames reconciled | PUBLISHED · REVIEW_ONLY |
| DFAST 2025 capital arithmetic | Federal Reserve, 22 banks | max error 0.000000 | PUBLISHED · REVIEW_ONLY |
| EBA 2025 capital arithmetic | EBA EU-wide, 64 banks | max error 0.000000 | PUBLISHED · REVIEW_ONLY |
| DebtRank replication | BIS LBS, EBA bank-sovereign, SEC 13F (4-quarter stability), FFIEC NIC (topology only) | six networks, receipts per sheet | PUBLISHED · REVIEW_ONLY |
| Crisis VaR/ES backtest ladder | 1,972 rows across public crises | 201 failures vs 49.3 expected – formal ES research rejects production validation | SELF-REJECTED · predates Aug 2026 kernel revision · re-run scheduled post-productionization |
| NYU V-Lab SRISK comparison | V-Lab method pages | not run | WITHHELD · license boundary |
| Own loss models (Level 3) | – | not run | WITHHELD · by decision |
Internal acceptance evidence
Internal acceptance evidence is separated from external replays by design. Estimated-config acceptance battery vs an internally generated 100M-path reference: 160/160 cells within the 5% envelope across 10 market regimes · median |error| 0.0156% · parameters hash-frozen before every comparison · every intermediate failure published · REVIEW_ONLY, pending production re-run · bundle sha256: a87246763088ca661e112e5b32ecd1fde4e80613e9cca9dc98f7afe19f2928c0. Details on the ARIN22 page.
All artifacts: github.com/saa-alliance/external-core-evidence, release v-r26. Verify: sha256sum -c SHA256SUMS.
Published research
Published research
The method is published under the founder’s name.
DRCET-1: Deterministic Risk-Compute Equivalence Test.
A Sampling-Resolution-Bounded Equivalence Benchmark for Institutional Risk-Compute Engines
Deterministic Risk-Compute Equivalence Testing
for Path-Dependent Functionals and Factor Attribution
Fail-Closed Release Gates for Numerical Operator Claims:
Certificates, Observed Order, and Spectral Domains in Risk Computation
Validating a Shared Deterministic Risk Kernel Across Two Peer Institutional Control Planes.
Backtest Evidence from Rolling-Window Validation and Cross-Product Implications
Low-Latency Stress Testing for AI-Integrated Macro-Financial Risk.
An Operator-Splitting Framework with Adaptive Error Calibration and Ontological Safety Controls
Adaptive Stress Testing via Operator Splitting.
A Multi-Operator Framework for Coupled Physical-Financial Risk Dynamics
Module-level methodology lives on each product page. Platform overview ->