PK · per-country backtest
Pakistan 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
163
since Apr 17
Accuracy @ 0.5
23.3%
38/163 correct
Brier score
0.567
lower is better
Observed positive rate
23%
mean predicted 64%
All 163 predictions (newest first)
| Eval date | Forecast | Pred ≥ 0.5? | Observed? | Correct? |
|---|---|---|---|---|
| Sep 26 | 80.1% | ↑ | no event | ✗ |
| Sep 25 | 78.4% | ↑ | no event | ✗ |
| Sep 24 | 95.0% | ↑ | no event | ✗ |
| Sep 23 | 78.1% | ↑ | no event | ✗ |
| Sep 22 | 80.4% | ↑ | no event | ✗ |
| Sep 21 | 95.0% | ↑ | no event | ✗ |
| Sep 20 | 95.0% | ↑ | no event | ✗ |
| Sep 19 | 78.5% | ↑ | no event | ✗ |
| Sep 18 | 70.4% | ↑ | no event | ✗ |
| Sep 17 | 78.2% | ↑ | no event | ✗ |
| Sep 16 | 81.7% | ↑ | no event | ✗ |
| Sep 15 | 74.7% | ↑ | no event | ✗ |
| Sep 14 | 74.8% | ↑ | no event | ✗ |
| Sep 13 | 61.3% | — | no event | ✓ |
| Sep 12 | 69.4% | ↑ | no event | ✗ |
| Sep 11 | 65.5% | ↑ | no event | ✗ |
| Sep 10 | 57.1% | — | no event | ✓ |
| Sep 9 | 67.4% | ↑ | no event | ✗ |
| Sep 8 | 65.7% | ↑ | no event | ✗ |
| Sep 7 | 78.7% | ↑ | no event | ✗ |
| Sep 6 | 75.2% | ↑ | no event | ✗ |
| Sep 5 | 72.3% | ↑ | no event | ✗ |
| Sep 4 | 76.9% | ↑ | no event | ✗ |
| Sep 3 | 72.8% | ↑ | no event | ✗ |
| Sep 2 | 72.5% | ↑ | no event | ✗ |
| Sep 1 | 87.4% | ↑ | no event | ✗ |
| Aug 31 | 86.9% | ↑ | no event | ✗ |
| Aug 30 | 88.7% | ↑ | no event | ✗ |
| Aug 29 | 78.0% | ↑ | no event | ✗ |
| Aug 28 | 72.3% | ↑ | no event | ✗ |
| Aug 27 | 68.3% | ↑ | no event | ✗ |
| Aug 26 | 69.8% | ↑ | no event | ✗ |
| Aug 25 | 72.5% | ↑ | no event | ✗ |
| Aug 24 | 72.1% | ↑ | no event | ✗ |
| Aug 23 | 82.1% | ↑ | no event | ✗ |
| Aug 22 | 68.0% | ↑ | no event | ✗ |
| Aug 21 | 67.7% | ↑ | no event | ✗ |
| Aug 20 | 67.9% | ↑ | no event | ✗ |
| Aug 19 | 64.4% | — | no event | ✓ |
| Aug 18 | 62.5% | — | no event | ✓ |
| Aug 17 | 62.1% | — | no event | ✓ |
| Aug 16 | 65.3% | ↑ | no event | ✗ |
| Aug 15 | 60.1% | — | no event | ✓ |
| Aug 14 | 68.6% | ↑ | no event | ✗ |
| Aug 13 | 57.7% | — | no event | ✓ |
| Aug 12 | 64.1% | — | no event | ✓ |
| Aug 11 | 68.0% | ↑ | no event | ✗ |
| Aug 10 | 55.9% | — | no event | ✓ |
| Aug 9 | 68.9% | ↑ | no event | ✗ |
| Aug 8 | 70.7% | ↑ | no event | ✗ |
| Aug 7 | 63.5% | — | no event | ✓ |
| Aug 6 | 75.5% | ↑ | no event | ✗ |
| Aug 5 | 68.6% | ↑ | no event | ✗ |
| Aug 4 | 69.7% | ↑ | no event | ✗ |
| Aug 3 | 80.7% | ↑ | no event | ✗ |
| Aug 2 | 78.9% | ↑ | no event | ✗ |
| Aug 1 | 78.0% | ↑ | no event | ✗ |
| Jul 31 | 76.6% | ↑ | no event | ✗ |
| Jul 30 | 75.8% | ↑ | no event | ✗ |
| Jul 29 | 73.0% | ↑ | 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/pk — current 7-day calibrated forecast for Pakistan with SHAP drivers
- /pk — Pakistan 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 Pakistan, plotted against what actually happened. 163 (predicted, observed) pairs from the rolling 30-day evaluation window.