Development notes
Honest writeups of what we shipped, what we measured, and what the numbers actually mean. Not a product blog. Not marketing copy. Closer to a lab notebook that happens to be public.
- · methodology · ML · graduation
Foreign Intervention, graduated: what eighteen years of indicator trajectories say
QGI's first ML-validated recipe earns its risk score. Four indicators, eighteen years, two hard validation gates cleared. The mechanic is similarity to a historical pre-crisis trajectory — and the line between similarity and probability is where this piece spends most of its time.
- · methodology · V2.0
What we mean when we say 'Recipe'
The word 'Recipe' on our methodology page has been doing more rhetorical work than the code. Here is the gap, why it opened, and what the V2.0 rebuild changes.
- · methodology · curation
Curation in practice: Iran, Ukraine, Venezuela
Three countries, three different gaps in our existing event coverage, three parallel curation passes in one afternoon. What we changed, what we didn't, and why this work sits upstream of every score we publish.
- · methodology · simulation · V1.9.2
Cross-country normalization meets the data
We proposed replacing per-country z-score normalization with a cross-country global approach. An Athena-backed simulation surfaced three problems before we shipped. Here's the finding, and why we're shelving V1.9.2 until event-curation coverage grows.
- · methodology · backtest · V1.9
The margin keeps widening
V1.8 → V1.8.1 → V1.9 — the gap between QGI and the naïve baseline has grown at every release. Here's why, and what the number does and doesn't prove.