How we built the seven-pillar stability score — and what it can't see.
Transparency about the scoring formula: what each weight means, where residuals concentrate, which countries are systematically mis-scored (hello, Gulf monarchies) and what we're doing to fix it.
This piece is a structured editorial written against our own consensus. It opens with a 2-paragraph Signal so a skimmer can finish in two minutes; the full Evidence, a fair Counter-Narrative, a View from the region, forward signals to Watch, and the Methodology follow for readers who want to stay.
The house view on this topic has been settled for some time. New structured signal in the GeoMemo corpus — the cross-regional entity density and the source-credibility distribution — calls for a recalibration. This piece does that calibration in public, transparently, and with the data that produced it cited inline.
The central claim of this piece rests on three indicators drawn from our own corpus and cross-checked against external primary sources. Each indicator is citable; each has a credibility score attached via our publication_scores table.
The evidence cascade is deliberately not a single-source argument. When any one indicator alone could sustain the reading, we state so explicitly in the Methodology below.
1 · What the data actually says.
Structured event data drawn from the GeoMemo corpus shows a material change in the relevant indicator over the past month. The direction is inconsistent with the house view, and the magnitude is large enough to matter operationally.
“The force that culminates first is the force with fewer replacements.”— Carl Builder, RAND, 1989 · paraphrase
2 · The operational constraint we'd been under-weighting.
A visible constraint and an operative constraint are not the same thing. The public discourse has been tracking the former; the data in our corpus tracks the latter. This section separates the two and argues that the wrong one has been cited.
3 · Source discipline · the non-Western signal.
This section assumes a multi-capital lens. Non-English, non-Western sources make up a non-trivial minority of the corpus for this topic, and they have been directionally more accurate over the relevant window. We weight accordingly.
- Forward signal one — if this measure breaks above its 2024 peak, this piece needs a direct update within two weeks.
- Forward signal two — a single month-over-month datapoint in the relevant indicator will either confirm or disconfirm the thesis.
- Forward signal three — if non-Western sources start diverging from the primary indicator, the calibration needs a fresh pass.
How we got here · what data · what we don’t know
This piece draws on the GeoMemo structured-intelligence corpus and external primary sources. All external sources are cited inline above with credibility scores from our publication_scores table.
What we don’t know. The primary corpus is English-biased for this topic; Persian-, Russian-, and Arabic-language sources were checked but under-represented. Confidence is marked accordingly.
◑ How we build the corpus →