ctianalyst.com

Source grading for analysts

A reliable source can report something false. An unreliable source can be right. Grading them on one axis loses the information that matters.

Source evaluation is the least glamorous tradecraft skill and one of the most frequently probed, because it is easy to check whether someone actually does it.

The Admiralty scale

Also called the NATO system, it grades every piece of reporting twice — once for the source, once for the information. A rating looks like B2.

Source reliabilityInformation credibility
A — Completely reliable1 — Confirmed by other sources
B — Usually reliable2 — Probably true
C — Fairly reliable3 — Possibly true
D — Not usually reliable4 — Doubtful
E — Unreliable5 — Improbable
F — Reliability cannot be judged6 — Truth cannot be judged

Reliability is a property of the source, built from its track record, methodology, access and incentives. Credibility is a property of this specific claim, judged mainly on whether independent reporting corroborates it and whether it is internally consistent.

The pairing is what carries information. An A6 — a highly reliable source making a claim nobody can corroborate — is a genuinely common and genuinely awkward situation, and the notation forces you to notice it.

The trap

The two axes are meant to be independent, and in practice analysts let reliability leak into credibility: a claim gets graded as more credible because a trusted vendor made it, rather than because anything corroborates it. Interviewers ask about this directly. Name the failure mode.

Circular reporting

The most common way corroboration is faked. Four articles reference a claim; three of them cite the fourth. You have one source, not four, and your confidence should reflect that.

How to catch it:

  • Trace every claim to the earliest publication that presents original evidence, not the most recent one.
  • Check whether "independent" sources share collection — two vendors using the same telemetry partner, or the same sinkhole, are not independent.
  • Watch for identical phrasing, identical indicator lists, or the same typo propagating.
  • Distinguish a source that observed something from one that agrees with someone who observed it.

Grading the sources CTI actually uses

SourceWhat to weigh
Your own telemetry Highest reliability available, but scoped to your visibility. Absence of evidence is not evidence of absence — know your blind spots.
Vendor research blogs Often excellent primary evidence, with a commercial incentive to publish and to emphasise novelty. Grade the evidence shown, not the brand. Reports that show their working deserve higher credibility than ones that assert conclusions.
Government advisories Usually high reliability and frequently based on non-public collection you cannot verify. Note that you are trusting rather than corroborating.
Sharing communities (ISACs, trust groups) High relevance to your sector. Reliability varies by contributor, and handling restrictions apply — see TLP.
Commercial feeds Ask how it was collected before asking how big it is. Volume is not quality, and unexplained provenance is a low grade regardless of price.
Criminal forums and leak sites Actors lie, exaggerate and re-list old data. Treat claims as claims; corroborate against victim-side evidence before reporting a breach as fact.
Social media research Fast and occasionally first. Frequently wrong in the first hours of an incident, and heavily prone to circular amplification.
Carry it through

Source grades are inputs to the confidence half of your assessment, not decoration in an appendix. If your judgement rests on C3 material, your confidence cannot be high, and your product should say why.

Practise source grading questions.

Next step

Run a mock interview on this material. You get the question, a graded response against a vetted rubric, and the probing follow-up an interviewer would ask next.

Related reading