The article A Better Crystal Ball. The Right Way to Think About the Future by J. Peter Scoblic and Philip E. Tetlock published on October 13, 2020 by Foreign Affairs proposes a reflection on how policymakers can improve their ability to anticipate and manage future crises, overcoming the limitations of traditional and often ineffective approaches.
Here is a summary of the key points:
1. Forecasting as an intrinsic part of policy-making Every policy decision is, after all, a forecast. However, geopolitical forecasting is notoriously difficult due to uncertainty, the lack of reliable data and the absence of universal laws governing the behaviour of international actors.
2. Decision-makers tend to project the past into the future, ignoring radical changes. As a result, they waste resources on outdated threats and are surprised by new ones, as happened with COVID-19.
3. Comparing two approaches: scenario planning and probabilistic forecasting Scenario planning originated in the context of the Cold War and consists of imagining plausible alternative futures to challenge dominant assumptions and stimulate cognitive flexibility.
Probabilistic forecasting, on the other hand, aims to quantify the risk of future events based on data, models and continuous updates. It is useful for short-term issues, but not very effective for long-term strategic decisions.
4. The limitations of both methods – Scenarios can be too broad and general to guide concrete action.
– Probabilistic forecasts, on the other hand, can be too narrow and technical and lack strategic relevance.
5. The proposal: integrate the two approaches through “clusters of questions”.
The authors proposes to develop clusters of precise questions that act as indicators that anticipate the realisation of certain scenarios. This makes it possible to
– link long-term visions to concrete short-term signals, – assign probabilities to future trajectories, – make political action more timely and informed.
6. Organisational and cultural barriers
The real obstacle is not theoretical but practical: the lack of will and institutional capacity to adopt these tools, coupled with the tendency of bureaucracies to focus on the immediate rather than the important.
7. The implicit conclusion
A combined approach – scenarios for imagining and probabilistic questions for measuring – could transform the way public policy is planned, providing decision-makers with better tools for dealing with uncertainty.
How to write a good intelligence report
In light of the article, a good intelligence report should be neither a theoretical treatise nor an anxiety-inducing narrative of the possible. Rather, it should be a bridge between future scenarios and present decisions. Below, I propose an ideal structure, enriched by reflections inspired by the text.
Structure of an intelligence report aimed at the political decision maker (according to the integrated scenario + probabilistic forecast model)
1. Executive Summary
• Maximum clarity, maximum impact.
• Two or three key (synthetic) scenarios for the medium/long-term future.
• For each: concrete indications of what is happening now that could signal its coming true.
• A “bottom line”: what should concern or reassure the decision maker today.
Example: “The ‘Triumphant China’ scenario for 2050 shows initial signs of materialization. The growth of soft power in Africa and territorial consolidation in the South China Sea indicate a trend that should not be ignored.”
2. Summary Scenario Planning
• Narrative description of 2–4 alternative, coherent, plausible but distinct scenarios.
• Each developed on a matrix of unknowns (e.g. economy + leadership, or technology + social instability).
• Objective: to open the decision maker’s mind, not to close off options.
The scenarios should not be prophecies, but well-constructed hypotheses to stimulate reflection and strategic preparation.
3. Cluster of forecasting questions
• For each scenario, indicate 3–5 specific questions (according to the “clairvoyance test” criterion: they must be able to receive a clear and verifiable answer).
• For each question: estimated probability (e.g. “40% chance of it happening within the next 12 months”), sources and degree of uncertainty.
Example:
• “By 2026, will China exercise operational control over Itu Aba?” ? Probability: 35%
• “In 2024, will CGTN overtake Voice of America for weekly audience in Nigeria?” ? Probability: 45%
4. Operational and strategic indications
• What can/should the decision maker do now to:
• reduce the risk of the most negative scenario,
• strengthen resilience with respect to multiple possibilities,
• seize emerging opportunities.
• Here intelligence must be bold: not just informing, but suggesting options (without prescribing).
Avoid generalities such as “monitor developments”. Better: “Arrange a targeted information campaign in Ghana within six months, to counter the Chinese narrative on the economic partnership”.
5. Updatability and feedback
• The report should include:
• a schedule of signals to be verified,
• a section for dynamic updating of probabilities,
• a feedback mechanism from the decision-maker and his/her staff (what was useful? what was not?).
Tone and language
• Clear but not simplistic. No academic jargon or abstract formulas.
• Avoid ideological judgments: the report should not convince, but enlighten.
• Take care of visualization: tables, matrices, predictive graphs can be very helpful.
In summary
A good intelligence report must answer three fundamental questions:
1. What could happen? (scenarios)
2. What signals are emerging now? (cluster of questions)
3. What can/should I do today, with this information? (practical decisions)
Follow