Method
Every CRI analysis follows the same method, so readers can see how a judgement was reached and how much weight it can bear.
Structured analytic techniques
Each analysis names the techniques it uses. The most common are these.
- Analysis of Competing Hypotheses (ACH). Several explanations are tested against the same evidence. The emphasis lies on evidence that contradicts a hypothesis, since evidence that fits every explanation proves little.
- Key Assumptions Check. The assumptions behind a judgement are listed and tested, so that a conclusion does not rest on something taken for granted.
- Indicators. Observable signs are defined in advance that would show a situation moving in one direction or another.
Explicit probability
Words such as “likely” mean different things to different readers. CRI therefore uses a fixed scale. When an analysis says an outcome is likely, it means a probability of 55 to 75 per cent, and the same word always carries the same meaning.
The probability scale
CRI uses the probability yardstick of the UK Professional Head of Intelligence Assessment. The gaps between ranges are deliberate and avoid false precision at the boundaries.
Confidence
Probability and confidence are kept apart. Probability states how likely an outcome is. Confidence states how solid the basis for that judgement is. CRI uses three levels.
- High. Good quality information from several independent sources.
- Moderate. Credible information, but with gaps or with sources that cannot be fully corroborated.
- Low. Fragmentary or doubtful information, or a judgement that rests heavily on assumptions.
Modelling
Where a question involves several uncertain factors, CRI uses Monte Carlo simulation. The model runs thousands of times with values drawn from defined ranges. The result shows the spread of possible outcomes rather than a single forecast. The assumptions and ranges are published with each model.
Sources
CRI uses open sources only. Each source is assessed for reliability and each piece of information for credibility, following the grading used in NATO intelligence practice. Every analysis lists its sources at the end.
Imagery and tracking data
- Satellite imagery. Optical and radar images from the European Copernicus programme show construction, activity at airfields and naval bases, and large vessels. Radar images also work through cloud and darkness, which matters in the Baltic and the Arctic. Each analysis states the resolution of the images used and what they can and cannot show.
- Ship and aircraft tracking. Ships and aircraft broadcast their positions openly. These signals show movements near undersea infrastructure, gaps where a vessel switches off its transmitter, and interference with satellite navigation.
- Open-source material. Official statements, state media, social media channels, procurement records and company registers, read in the original language where possible, including Russian.
Limits
CRI works only with information that is openly available. It does not intercept communications, does not track private individuals, and keeps its focus on the activities of states, their forces and the companies that serve them.