QUANTITATIVE IMMUNE SYSTEM
The Quantitative Immune System is one of ASTRIVERA's two core capabilities. It works across data, research, portfolio, execution and accounting components. Structured Evidence is used to check operational state, identify deviations and preserve the basis for diagnosis and recovery. Used independently, it provides local Read-only Diagnosis. When connected to the complete system, it extends into continuous immunity, an Independent Production Gate and historical immune memory.
The Quantitative Immune System is currently in research and internal validation. This page describes product direction, component structure and capabilities available for demonstration at the present stage. Diagnostic scope, supported environments, rule coverage, verification methods and gating authority depend on the current version and agreed scope.
Read authorised evidence within the client environment and produce a coverage statement, diagnostic findings and recommendations for further examination. The diagnostic process does not write to client production objects. Reports and temporary artifacts remain in the client environment, and the client decides whether they are exported.
Once connected to the complete system, record data, decision, execution, accounting and recovery states continuously. Production permission is administered by an Independent Production Gate, while historical issues and their treatment enter immune memory.
Record the components, paths, time period and content level authorised for reading. Protected scope remains outside the diagnostic view.
Read authorised data, configuration, research artifacts and operating records. Each item retains its source, time, version and permitted use.
Check material facts against the current rules and assembly configuration. Applicable rules, skipped rules and evidence gaps are recorded separately.
Compare expected state with actual results to locate the affected component and scope. Facts, inference and diagnostic confidence are presented separately.
When replay or a focused test is required, authorised inputs move into an isolated environment. Test results return to the diagnostic structure as new evidence.
After an issue is addressed, recheck material results against the same scope and rules. Recovery state is linked to review evidence, exceptions and approval records.
Examine missing or duplicated data, time misalignment, security identity and corporate-action handling. Confirm that relevant data was available at the decision time.
Examine data snapshots, strategy versions, configuration, parameters and dependencies. Determine whether a research run can be reconstructed under the same conditions.
Examine trading calendars, fill assumptions, fees, slippage and market restrictions. Confirm that the conditions underlying a backtest result were retained in full.
Reconcile model output, portfolio weights, cash and risk adjustments. Confirm that risk constraints changed the target itself rather than appearing only in a report.
Connect the order plan, broker responses, fills, positions, cash and profit and loss. Partial fills, cancellations, retries and unresolved states are handled separately.
Record incidents, treatment actions, exceptions, review results and recovery state. When a similar issue returns, connect it to historical evidence and prior treatment records.
