BY · per-country backtest
Belarus forecast vs reality
Compare published forecasts with observed outcomes for this country.
Updated every 30 min · CC BY 4.0 · Raw JSON · Current forecast →
Forecast time series
Blue line: forecast probability. Green ✓ markers: forecast was right. Red ✗ markers: forecast was wrong. Dashed line at 0.5 is the binary decision threshold.
Y axis: forecast probability · Faint tick: distance to observed outcome (0 or 1)
Forecasts evaluated
162
since Apr 17
Accuracy @ 0.5
34.0%
55/162 correct
Brier score
0.364
lower is better
Observed positive rate
13%
mean predicted 42%
All 162 predictions (newest first)
| Eval date | Forecast | Pred ≥ 0.5? | Observed? | Correct? |
|---|---|---|---|---|
| Sep 26 | 81.3% | ↑ | no event | ✗ |
| Sep 25 | 80.3% | ↑ | no event | ✗ |
| Sep 24 | 79.5% | ↑ | no event | ✗ |
| Sep 23 | 77.6% | ↑ | no event | ✗ |
| Sep 22 | 77.6% | ↑ | no event | ✗ |
| Sep 21 | 80.9% | ↑ | no event | ✗ |
| Sep 20 | 80.9% | ↑ | no event | ✗ |
| Sep 19 | 78.2% | ↑ | no event | ✗ |
| Sep 18 | 83.0% | ↑ | no event | ✗ |
| Sep 17 | 79.0% | ↑ | no event | ✗ |
| Sep 16 | 79.8% | ↑ | no event | ✗ |
| Sep 15 | 80.6% | ↑ | no event | ✗ |
| Sep 14 | 87.5% | ↑ | no event | ✗ |
| Sep 13 | 5.2% | — | no event | ✓ |
| Sep 12 | 70.8% | ↑ | no event | ✗ |
| Sep 11 | 58.7% | — | no event | ✓ |
| Sep 10 | 68.4% | ↑ | no event | ✗ |
| Sep 9 | 69.2% | ↑ | no event | ✗ |
| Sep 8 | 64.5% | — | no event | ✓ |
| Sep 7 | 68.5% | ↑ | no event | ✗ |
| Sep 6 | 66.0% | ↑ | no event | ✗ |
| Sep 5 | 69.4% | ↑ | no event | ✗ |
| Sep 4 | 64.9% | ↑ | no event | ✗ |
| Sep 3 | 59.1% | — | no event | ✓ |
| Sep 2 | 77.5% | ↑ | no event | ✗ |
| Sep 1 | 77.5% | ↑ | no event | ✗ |
| Aug 31 | 74.0% | ↑ | no event | ✗ |
| Aug 30 | 75.3% | ↑ | no event | ✗ |
| Aug 29 | 75.5% | ↑ | no event | ✗ |
| Aug 28 | 72.8% | ↑ | no event | ✗ |
| Aug 27 | 65.0% | ↑ | no event | ✗ |
| Aug 26 | 72.2% | ↑ | no event | ✗ |
| Aug 25 | 70.2% | ↑ | no event | ✗ |
| Aug 24 | 66.1% | ↑ | no event | ✗ |
| Aug 23 | 66.3% | ↑ | no event | ✗ |
| Aug 22 | 71.6% | ↑ | no event | ✗ |
| Aug 21 | 73.9% | ↑ | no event | ✗ |
| Aug 20 | 70.8% | ↑ | no event | ✗ |
| Aug 19 | 63.3% | ↑ | no event | ✗ |
| Aug 18 | 68.7% | ↑ | no event | ✗ |
| Aug 17 | 73.6% | ↑ | no event | ✗ |
| Aug 16 | 77.1% | ↑ | no event | ✗ |
| Aug 15 | 78.6% | ↑ | no event | ✗ |
| Aug 14 | 78.3% | ↑ | no event | ✗ |
| Aug 13 | 81.5% | ↑ | no event | ✗ |
| Aug 12 | 75.5% | ↑ | no event | ✗ |
| Aug 11 | 82.3% | ↑ | no event | ✗ |
| Aug 10 | 77.9% | ↑ | no event | ✗ |
| Aug 9 | 71.4% | ↑ | no event | ✗ |
| Aug 8 | 74.4% | ↑ | no event | ✗ |
| Aug 7 | 82.6% | ↑ | no event | ✗ |
| Aug 6 | 59.2% | — | no event | ✓ |
| Aug 5 | 71.8% | ↑ | no event | ✗ |
| Aug 4 | 77.2% | ↑ | no event | ✗ |
| Aug 3 | 73.5% | ↑ | no event | ✗ |
| Aug 2 | 68.7% | ↑ | no event | ✗ |
| Aug 1 | 66.9% | ↑ | no event | ✗ |
| Jul 31 | 70.3% | ↑ | no event | ✗ |
| Jul 29 | 44.6% | — | no event | ✓ |
| Jul 28 | 49.3% | ↑ | no event | ✗ |
How to read this
- Each row is one historical forecast. We made the prediction at eval_date, then waited the 7-day horizon, then graded against the observed outcome.
- Forecast % is the calibrated probability we published that day. Post-recalibration (2026-05-20) these now match actual observed rates much better — see the refit finding.
- Pred ≥ 0.5 shows the binary alert decision. Note we usually fire alerts at a lower threshold (see /v1/sentinel/global_heatmap for the live cutoff) — the 0.5 column here is for backtest scoring consistency with the global confusion matrix.
- Correct = both (pred ≥ 0.5) and (observed) agree.
Related
- /atlas/forecast/by — current 7-day calibrated forecast for Belarus with SHAP drivers
- /by — Belarus country page (current state, recent incidents)
- /sentinel/backtest — global reliability diagram + per-country comparison table
- Calibration refit writeup
About this view
Every Voidly Sentinel shutdown-risk forecast for Belarus, plotted against what actually happened. 162 (predicted, observed) pairs from the rolling 30-day evaluation window.