This is a dense period of overlapping strategic risk, and most serious assessment of it happens out of public view. SIGMA keeps an open, committed record of where the world's active conflicts are heading, and publishes it to support transparency and informed public discussion. For deeper institutional-level work, a fuller picture of shared risk is available through membership and commissioned work.
Every active dyad on one map, sized and coloured by kinetic risk — the full picture at a glance. Each is decomposed four ways, committed to a time-stamped ledger before it resolves, and scored against what happens.
Open the tracker →The ledger has just opened. Every forecast is scored only as it resolves — the record here begins at the first resolution and is never backfilled or revised after the fact. Until then the score stands empty, on purpose.
Follow SIGMA's forecasting record.
By signing up, you agree to receive communications from SIGMA Futures.
We respect your privacy and you can opt out at any time.
Each dyad forecast carries SIGMA's current committed assessment, decomposed and dated. The headline figure is the probability of kinetic conflict within the horizon; the bar beneath divides the assessment four ways, from negotiated settlement through to major escalation. Together, they show the range of possible futures inherent to each forecast.
SIGMA forecasts begin with a structured assessment of the forces shaping each confrontation, including incentives, capabilities, bargaining conditions, political constraints, escalation pathways, and the signals most likely to change them.
These inputs are integrated within a common forecasting framework and tested against the strongest case for a different outcome before a probability is committed. The reasoning behind each probability therefore remains explicit and open to subsequent evaluation.
Yes. AI extends SIGMA's research, synthesis, and drafting capacity, particularly where large volumes of conflicting evidence must be compared across multiple active confrontations.
Human judgement remains responsible for the framework, the interpretation, and the forecast itself. Probabilities are committed by SIGMA, timestamped in the forecasting ledger, and scored against subsequent outcomes. AI helps us see more; it does not inherit responsibility for what we conclude.
A committed forecast does not change. The assessment can.
Once committed, a forecast remains in the ledger with its original probability, horizon, and date. New evidence may change SIGMA's current view, but it does not rewrite the earlier record. Material changes are documented through subsequent assessments and interim signal updates, while the original commitment remains available for scoring.
This separation allows the analysis to remain responsive without making the track record retrospective.
Forecasts about armed conflict abound. Any serious analytical body offers its own map of looming flashpoints, and audiences encounter no shortage of emphatic claims. What is genuinely scarce is verifiable accountability.
The forecasting landscape
| Function | Primary question | Typical output | Representative organisations | Public accountability |
|---|---|---|---|---|
| Reporting | What happened? | Verified events | Reuters, AP | Verified on publication |
| Event data | What has been recorded? | Structured records | ACLED, UCDP | Historical record |
| Strategic analysis | What does it mean? | Interpretation | ISW, RAND | Through reputation |
| Scenario analysis | What could happen? | Plausible futures | RANE, Eurasia Group | Constrained by structure |
| Probabilistic forecasting | How likely is it? | Explicit probabilities | Metaculus, Good Judgment | Outcome can be checked |
| Forecast accountability | Were the forecasts right? | A scored record | — | Public scoring by design |
Most geopolitical reporting and analysis informs judgement. SIGMA commits one. Every assessment records an explicit probability before the outcome, together with the reasoning behind it and the indicators most likely to alter the outlook. Assessed against a consistent framework, active confrontations can be compared rather than followed in isolation.
Committed assessments create a different kind of signal — one that can be challenged while it stands, superseded on the record when the evidence changes, and scored once the forecasting horizon closes. When you have to defend a call, that consistency matters.
Qualitative analysis can describe situations with markedly different risk profiles in similar terms — “elevated,” “concerning,” “tensions remain high.” A probability adds resolution that vocabulary alone lacks. For instance, a forecast of 70% does not mean that an event will occur; it judges the event substantially more likely than not. A forecast of 30% is a materially different judgement, though both might still be described as concerning. That precision comes with a cost. A committed probability can be proved wrong, where “elevated” cannot.
SIGMA does not claim that geopolitical events are predictable with certainty. Each forecast commits to a probability before the outcome is known, defines in advance what would count as the event occurring, and leaves a public record. Over time, that record allows a stricter test: whether outcomes assigned higher probabilities occur more often than those assigned lower ones.
SIGMA maintains a committed forecasting ledger as the foundation of its assessments. Where appropriate, forecasts are published openly to promote transparency, accountability, and informed public discussion. Membership, commissioned work, and voluntary support sustain that work. If you value independent, transparent geopolitical forecasting, you can help keep the public record open.
Support the Public RecordStructured forecasting can help reveal where systems appear most sensitive to change, making it easier to distinguish interventions likely to be consequential from those that merely respond to symptoms. Done well, it progressively uncovers the relationships that make some futures more likely than others, clarifying where different choices still matter.
Once those relationships are better understood, the question naturally shifts from prediction to participation: which of these relationships are fixed constraints, and which remain open to influence? SIGMA's orientation toward forecasting is not, finally, to demonstrate foresight, but to help widen the space in which wiser collective action remains possible.
A risk of forecasting is that as strategies accumulate, uncertainty can seem increasingly tamed. Yet if every scenario quietly converges on the same worldview, planning ceases to explore possibility and begins reproducing inevitability. The result is not resilience, but self-enclosure.
Every forecast is ultimately a hypothesis about structure rather than an assertion about destiny. Even highly probable futures remain embedded within living systems that can still learn, adapt, and — under the right conditions — change course.