Every score =
severity weight × confidence weight.
| Severity | Weight |
|---|
| High | 10 |
| Medium | 6 |
| Context | 3 |
| Confidence status | Weight |
|---|
| Confirmed | 1.0 |
| Developing | 0.75 |
| Pending verification | 0.6 |
| Projection | 0.5 |
Important: these weights are an editorial judgment call made for internal consistency — they are not statistically derived, externally validated, or based on actuarial/loss data. A different analyst could reasonably pick different numbers (e.g. 0.8 instead of 0.75 for "Developing"). What makes this defensible isn't that the weights are "correct" — it's that the same weights are applied to every item, every cycle, and shown here in full rather than hidden behind an opaque score.
Today's Signal Strength = average score of items currently in the Hot zone (max 4, by design).
Cycle average = average score of all
unique signals this cycle (17 cards shown, 1 is a cross-tab duplicate of the same underlying Interpol finding, counted once → 16 unique signals).
Risk Radar categories = average score of items tagged to that category (typically 3–4 items each this cycle — small samples, treat as directional). 2 items (NSE AI ETF, open-weight models) don't fit any category and are excluded, not force-fit.
No number on this page uses invented precision (e.g. a fake "94% confidence") — confidence is always shown as one of four disclosed labels, and every score can be traced back to the item(s) that produced it.