JANTARJantar Risk: Evidence, Not Estimates
A cited record of what the spacecraft actually did, sealed so that a change to it shows. Observed events are kept strictly separate from projections, and the carrier decides.
Objective Telemetry Evidence Over Adjusted Estimates
What Happened
- Every deviation over the policy period, each carrying the citations of the investigation that found it.
- Findings the evidence does not support are shown as unsupported, never removed.
- Each pre-agreed trigger: whether the condition was met, and the observed value it was tested on.
What Might Happen
- Optional, and off by default.
- Cites the model, version and configuration behind anything it projects.
- Never merged with the observed record into one attestation.
What the package contains
Jantar Risk does not score a spacecraft out of 100 and does not publish a probability of constructive total loss. No underwriter prices on a vendor’s probability, and constructive total loss in a real policy is defined on transponder-unit-years (capacity × life), not on a bus health score. What we can produce, and an operator cannot easily assemble by hand, is the evidence.
| Section | What it is | Status |
|---|---|---|
| Observed-anomaly ledger | Every deviation over the policy period, each carrying the citations of the investigation that found it. The renewal health report. | Statement of fact, sealed |
| Unsupported findings | Findings the evidence does not support are shown as unsupported, never removed. | Disclosure, within the ledger seal |
| Projection annex | Optional and off by default. Kept out of the observed record, and cites the model, version and configuration behind anything it projects. | Forecast, not fact. Sealed separately, never merged with the observed seal |
| Trigger evaluation | Whether each pre-agreed policy condition was met, and the observed value it was tested on. | Input to the carrier’s decision, sealed |
| Disaster-scenario exposure | Which policies in a carrier-supplied schedule each Lloyd’s Realistic Disaster Scenario touches, and the gross loss under that scenario’s own terms. It needs no telemetry, and it reports no total, because the scenarios are alternatives. | Arithmetic on the carrier’s own schedule, sealed |
Survival priors ship as published literature and are labelled reference-uncalibrated; nothing is calibrated to a customer fleet until a licensed dataset is in place (ADR 0021). Jantar prepares evidence and never files, prices, settles or authorises a payment.
Verifiable, Not Asserted
An underwriter can re-compute a seal and see that the record has not changed. Catching a rewrite of the whole chain needs its head kept outside our database: the hosted service writes the chain head to Cloud Storage every hour. That bucket has no retention lock yet, and a self-hosted deployment anchors only once it turns anchoring on.