Multi-horizon forecast index
Three honest forecast horizons per country — each with its own XGBoost model, its own SHAP attribution, and its own 90% conformal interval. Click a country for per-horizon detail.
Trained 2026-05-21 · Recommendation: ship_all_three · model info
20 spotlight countries · ranked by 30-day risk
| # | Country | 1d | 7d | 30d | 30d range (90%) | Cons. |
|---|---|---|---|---|---|---|
| 1 | Iran IR | 22% | 36% | 95% | [45, 100] | ok |
| 2 | China CN | 22% | 36% | 95% | [45, 100] | ok |
| 3 | Azerbaijan AZ | 27% | 59% | 91% | [41, 100] | ok |
| 4 | Uzbekistan UZ | 27% | 59% | 91% | [48, 100] | ok |
| 5 | India IN | 22% | 59% | 91% | [41, 100] | ok |
| 6 | Egypt EG | 36% | 59% | 83% | [33, 100] | ok |
| 7 | Venezuela VE | 36% | 36% | 83% | [33, 100] | ok |
| 8 | Myanmar MM | 36% | 36% | 83% | [42, 100] | ok |
| 9 | Pakistan PK | 36% | 59% | 83% | [33, 100] | ok |
| 10 | Turkey TR | 36% | 36% | 83% | [33, 100] | ok |
| 11 | Russia RU | 22% | 36% | 70% | [20, 100] | ok |
| 12 | Belarus BY | 15% | 59% | 70% | [29, 100] | ok |
| 13 | Cuba CU | 15% | 36% | 70% | [20, 100] | ok |
| 14 | Ethiopia ET | 22% | 49% | 70% | [20, 100] | ok |
| 15 | Turkmenistan TM | 22% | 36% | 70% | [20, 100] | ok |
| 16 | TJ TJ | 27% | 59% | 70% | [20, 100] | ok |
| 17 | Kazakhstan KZ | 27% | 59% | 70% | [20, 100] | ok |
| 18 | AF AF | 27% | 59% | 70% | [20, 100] | ok |
| 19 | Bangladesh BD | 27% | 59% | 70% | [20, 100] | ok |
| 20 | Saudi Arabia SA | 15% | 59% | 70% | [20, 100] | ok |
Cons. = monotonicity (P(1d) ≤ P(7d) ≤ P(30d)). 30d range = 90% conformal interval.
Why three horizons?
A single 7-day number is what most journalists ask for, but it hides two important things: what is going to happen tomorrow, and what the operational ceiling is over the next month. We train three independent XGBoost models and publish all three with honest LOCO AUC numbers and per-horizon SHAP attribution.
The monotonicity check is a free honesty signal: longer windows must contain shorter windows, so if P(1d) > P(30d) the three models are disagreeing and you should treat the headline numbers with caution. See /atlas/models for the full registry, or /methodology for the full pipeline.
Related
- /atlas/forecast — calibrated 7-day forecast index
- /atlas/score-v2 — level-aware composite score (uses 30d avg)
- /atlas/changelog — model registry history
- /sentinel/calibration — live model-honesty dashboard