ER · per-country backtest
ER forecast vs reality
Every Voidly Sentinel shutdown-risk forecast for ER, plotted against what actually happened. 112 (predicted, observed) pairs from the rolling 30-day evaluation window.
Updated every 30 min · CC BY 4.0 · Raw JSON · Current forecast →
Forecasts evaluated
112
since Apr 17
Accuracy @ 0.5
77.7%
87/112 correct
Brier score
0.002
lower is better
Observed positive rate
0%
mean predicted 4%
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)
All 112 predictions (newest first)
| Eval date | Forecast | Pred ≥ 0.5? | Observed? | Correct? |
|---|---|---|---|---|
| Aug 6 | 3.7% | — | no event | ✓ |
| Aug 5 | 3.8% | — | no event | ✓ |
| Aug 4 | 3.0% | — | no event | ✓ |
| Aug 3 | 4.8% | — | no event | ✓ |
| Aug 2 | 3.0% | — | no event | ✓ |
| Aug 1 | 5.1% | ↑ | no event | ✗ |
| Jul 31 | 4.1% | — | no event | ✓ |
| Jul 30 | 4.0% | — | no event | ✓ |
| Jul 29 | 4.7% | — | no event | ✓ |
| Jul 28 | 4.0% | — | no event | ✓ |
| Jul 27 | 3.7% | — | no event | ✓ |
| Jul 26 | 4.2% | — | no event | ✓ |
| Jul 25 | 5.1% | ↑ | no event | ✗ |
| Jul 24 | 5.3% | ↑ | no event | ✗ |
| Jul 23 | 4.9% | — | no event | ✓ |
| Jul 22 | 3.0% | — | no event | ✓ |
| Jul 21 | 5.4% | ↑ | no event | ✗ |
| Jul 20 | 5.3% | ↑ | no event | ✗ |
| Jul 19 | 4.3% | — | no event | ✓ |
| Jul 18 | 4.6% | — | no event | ✓ |
| Jul 17 | 3.3% | — | no event | ✓ |
| Jul 16 | 3.8% | — | no event | ✓ |
| Jul 15 | 4.3% | — | no event | ✓ |
| Jul 14 | 5.4% | ↑ | no event | ✗ |
| Jul 13 | 3.6% | — | no event | ✓ |
| Jul 12 | 3.8% | — | no event | ✓ |
| Jul 11 | 5.1% | ↑ | no event | ✗ |
| Jul 10 | 3.5% | — | no event | ✓ |
| Jul 9 | 4.9% | — | no event | ✓ |
| Jul 8 | 4.7% | — | no event | ✓ |
| Jul 7 | 5.3% | ↑ | no event | ✗ |
| Jul 6 | 4.9% | — | no event | ✓ |
| Jul 5 | 4.7% | — | no event | ✓ |
| Jul 4 | 4.0% | — | no event | ✓ |
| Jul 3 | 4.4% | — | no event | ✓ |
| Jul 2 | 5.0% | ↑ | no event | ✗ |
| Jul 1 | 5.3% | ↑ | no event | ✗ |
| Jun 30 | 4.6% | — | no event | ✓ |
| Jun 29 | 4.9% | — | no event | ✓ |
| Jun 28 | 3.2% | — | no event | ✓ |
| Jun 27 | 4.2% | — | no event | ✓ |
| Jun 26 | 4.9% | — | no event | ✓ |
| Jun 25 | 5.0% | ↑ | no event | ✗ |
| Jun 24 | 4.4% | — | no event | ✓ |
| Jun 23 | 5.3% | ↑ | no event | ✗ |
| Jun 22 | 2.5% | — | no event | ✓ |
| Jun 21 | 3.1% | — | no event | ✓ |
| Jun 20 | 4.5% | — | no event | ✓ |
| Jun 19 | 4.2% | — | no event | ✓ |
| Jun 18 | 3.1% | — | no event | ✓ |
| Jun 17 | 4.1% | — | no event | ✓ |
| Jun 16 | 5.2% | ↑ | no event | ✗ |
| Jun 15 | 4.8% | — | no event | ✓ |
| Jun 14 | 5.4% | ↑ | no event | ✗ |
| Jun 13 | 3.5% | — | no event | ✓ |
| Jun 12 | 3.2% | — | no event | ✓ |
| Jun 11 | 5.1% | ↑ | no event | ✗ |
| Jun 10 | 4.1% | — | no event | ✓ |
| Jun 9 | 4.7% | — | no event | ✓ |
| Jun 8 | 3.5% | — | 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/er — current 7-day calibrated forecast for ER with SHAP drivers
- /er — ER country page (current state, recent incidents)
- /sentinel/backtest — global reliability diagram + per-country comparison table
- Calibration refit writeup