What Geostrategic Watch Indicators Actually Do in a Live Risk Process

In 1964, Sherman Kent published an essay in the CIA's classified journal Studies in Intelligence showing that the word "probable," when used in intelligence assessments, was being read by different decision-makers as meaning anywhere between 30 and 75 percent likelihood.¹ Kent proposed a numerical probability framework to close that gap. The intelligence community did not adopt it, and most commercial geopolitical risk processes today carry the same flaw built into every indicator list they maintain.

Why geostrategic watch lists accumulate entries without triggering decisions

Most corporate watch lists describe conditions rather than triggering thresholds, which means there is no mechanism for converting a watched situation into a decision point. The entry stays on the list, the note says "tensions remain elevated," and the review cycle repeats.

• Language without commitment. Kent's 1964 survey found that the same estimative terms produced radically different probability interpretations across readers.¹ Without attached probability estimates, risk reports describe without committing. Different decision-makers draw different conclusions from the same material and proceed on different assumptions.
• No trigger architecture. RAND Corporation's Scalable Warning and Resilience Model identifies the structural failure directly: "In slowly developing situations it is often hard to stimulate action."² An indicator without a threshold and a linked decision step is a monitoring record, not an operational tool.
• Accumulation without resolution. Watch lists grow as analysts add items to demonstrate coverage. They rarely shrink, because ambiguity makes closure hard to confirm. After several quarterly reviews a list contains thirty or forty entries. Decision-makers scan it, note that everything remains under observation, and move on.
• Committee dilution. When a geopolitical risk indicator appears in an executive briefing without a probability estimate and without a linked decision threshold, it becomes one item competing for attention among quarterly financials and operational KPIs. It is noted and the meeting continues.

How confirmation bias compounds the failure over time

A second failure mechanism works inside the analyst reviewing the indicator. Once a view is formed about a developing situation, new information tends to be processed as confirming that picture rather than genuinely updating it.

• Kahneman on evidence processing. Daniel Kahneman's research on System 1 and System 2 thinking documents that initial impressions shape how subsequent information is processed.⁴ Evidence that contradicts a stable picture requires deliberate cognitive effort to take seriously. Most structured review processes do not build in that effort.
• Pre-mortem as counter. The pre-mortem technique asks the analyst to assume the worst-case outcome has already occurred and work backwards to what conditions would have had to hold for it to happen. That exercise forces engagement with the disconfirming evidence that confirmation bias would otherwise suppress.⁴
• RAND SWARM protocol. RAND's SWARM framework requires each indicator to be paired with an alternative explanation, a set of assumptions that must be tested before the indicator is accepted as confirmed, and an explicit account of what observable evidence would disconfirm the current assessment.⁵ The discipline forces analysts to confront uncertainty rather than paper over it with cautious language.
• 2025 research finding. A 2025 paper in Comparative Strategy concluded that the strategic warning discipline Kent helped establish requires revival because the institutional pressures producing ambiguous language persist regardless of analytical tools available to practitioners.³

What calibrated geostrategic watch indicators require

A watch indicator that functions as an operational tool rather than a monitoring record has four components that most corporate indicator lists do not include.

• Named observable condition. The indicator specifies an observable event or data point, not a general situation. "VLCC voyage rerouting via Lombok Strait exceeds fifteen departures per rolling seven-day period, as tracked in commercial AIS data" is observable and verifiable. "Monitor maritime tensions" is neither.
• Defined threshold. The indicator specifies what change in the observable condition crosses into the action zone. Without a threshold, analysts can always justify continued monitoring rather than escalating to a decision. The threshold converts watching into deciding.
• Linked decision step. Each threshold crossing maps to a specific decision: a 48-hour routing cost review, a contract force majeure clause assessment, a sourcing alternative evaluation. Without that link, the indicator has nowhere to go when it activates and the watching continues regardless.
• Named owner. Someone is responsible for monitoring the observable condition and escalating when the threshold is reached. Unowned indicators are reviewed collectively and acted on by no one specifically. Ownership converts a shared awareness into an operational accountability.
• Probability estimate. Kent's recommendation from 1964 still holds: attaching a stated probability to the risk condition creates an expectation that can be tracked, updated, and tested against subsequent evidence. It also forces the analyst to commit to a view rather than hedge across all possible outcomes.

Meridian Intell note: Meridian structures its geostrategic risk monitoring around calibrated watch indicators linked to client-defined decision thresholds. Operational alerting cycles are matched to the lead times required for the decisions each indicator supports.

Methodology: Meridian Intell field notes draw on primary research, peer-reviewed scholarship, and practitioner analysis. Sources are cited for verifiability. Where analysis goes beyond available evidence, it is identified as such.

Footnotes

1 Sherman Kent, "Words of Estimative Probability," Studies in Intelligence, Vol. 8, No. 4 (Fall 1964), pp. 49-65. Declassified and publicly available via the CIA's Center for the Study of Intelligence, https://www.cia.gov/static/Words-of-Estimative-Probability.pdf. Kent documented variability in how analysts and readers interpreted probability terms across a survey of intelligence community personnel.

2 RAND Corporation, "Scalable Warning and Resilience Model (SWARM): Threat Warning in a Changing World," RAND Project Air Force, PE-383-A/OSD, 2021. The passage cited appears in the executive summary discussion of slowly developing threats. Available at https://www.rand.org/pubs/perspectives/PE383.html.

3 John R. Deni and Kathleen J. McInnis, "Strategic Warning Intelligence: Revival Needed," Comparative Strategy, Vol. 44, No. 2 (2025), pp. 112-128. The paper documents the atrophying of dedicated warning intelligence functions in both government and private sector intelligence operations.

4 Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011). System 1 and System 2 framework as applied to evidence processing and the pre-mortem technique discussed in the chapters on overconfidence and planning.

5 RAND Corporation, SWARM framework documentation, ibid. Core elements: named indicators linked to event drivers; assumption testing protocols; structured consideration of alternative explanations; escalation thresholds defined in advance. See also Kristan J. Wheaton and Michael T. Beerbower, "Towards a New Definition of Intelligence," Stanford Law and Policy Review, Vol. 17, No. 1 (2006), for the underlying epistemological framework.

About the author

Shekhar Attri, Co-Founder & CTO. An Indian Army Special Forces veteran with 21 years of service and a gallantry medal, Shekhar's corporate security advisory work spans Singapore, India, the Philippines, and the UAE, alongside PhD research on machine intelligence under incomplete information.