CN · per-country backtest
China forecast vs reality
Every Voidly Sentinel shutdown-risk forecast for China, plotted against what actually happened. 73 (predicted, observed) pairs from the rolling 30-day evaluation window.
Updated every 30 min · CC BY 4.0 · Raw JSON · Current forecast →
Forecasts evaluated
73
since Apr 17
Accuracy @ 0.5
39.7%
29/73 correct
Brier score
0.005
lower is better
Observed positive rate
0%
mean predicted 6%
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 73 predictions (newest first)
| Eval date | Forecast | Pred ≥ 0.5? | Observed? | Correct? |
|---|---|---|---|---|
| Jun 28 | 2.4% | — | no event | ✓ |
| Jun 27 | 2.1% | — | no event | ✓ |
| Jun 26 | 2.6% | — | no event | ✓ |
| Jun 25 | 1.7% | — | no event | ✓ |
| Jun 24 | 9.6% | ↑ | no event | ✗ |
| Jun 23 | 9.0% | ↑ | no event | ✗ |
| Jun 22 | 4.8% | — | no event | ✓ |
| Jun 21 | 3.6% | — | no event | ✓ |
| Jun 20 | 5.0% | ↑ | no event | ✗ |
| Jun 19 | 4.7% | — | no event | ✓ |
| Jun 18 | 3.5% | — | no event | ✓ |
| Jun 17 | 4.3% | — | no event | ✓ |
| Jun 16 | 5.2% | ↑ | no event | ✗ |
| Jun 15 | 4.9% | — | no event | ✓ |
| Jun 14 | 4.8% | — | no event | ✓ |
| Jun 13 | 3.8% | — | no event | ✓ |
| Jun 12 | 3.5% | — | no event | ✓ |
| Jun 11 | 4.8% | — | no event | ✓ |
| Jun 10 | 4.6% | — | no event | ✓ |
| Jun 9 | 3.6% | — | no event | ✓ |
| Jun 8 | 5.4% | ↑ | no event | ✗ |
| Jun 7 | 4.9% | — | no event | ✓ |
| Jun 6 | 4.9% | — | no event | ✓ |
| Jun 5 | 5.0% | ↑ | no event | ✗ |
| Jun 4 | 10.0% | ↑ | no event | ✗ |
| Jun 3 | 9.6% | ↑ | no event | ✗ |
| Jun 2 | 12.1% | ↑ | no event | ✗ |
| Jun 1 | 12.8% | ↑ | no event | ✗ |
| May 31 | 10.7% | ↑ | no event | ✗ |
| May 30 | 14.3% | ↑ | no event | ✗ |
| May 29 | 13.5% | ↑ | no event | ✗ |
| May 28 | 15.3% | ↑ | no event | ✗ |
| May 27 | 11.0% | ↑ | no event | ✗ |
| May 26 | 9.6% | ↑ | no event | ✗ |
| May 25 | 8.2% | ↑ | no event | ✗ |
| May 24 | 8.3% | ↑ | no event | ✗ |
| May 23 | 7.1% | ↑ | no event | ✗ |
| May 22 | 6.6% | ↑ | no event | ✗ |
| May 21 | 9.2% | ↑ | no event | ✗ |
| May 20 | 8.2% | ↑ | no event | ✗ |
| May 19 | 6.5% | ↑ | no event | ✗ |
| May 18 | 9.0% | ↑ | no event | ✗ |
| May 17 | 5.6% | ↑ | no event | ✗ |
| May 16 | 7.5% | ↑ | no event | ✗ |
| May 15 | 5.0% | ↑ | no event | ✗ |
| May 14 | 6.5% | ↑ | no event | ✗ |
| May 13 | 6.6% | ↑ | no event | ✗ |
| May 12 | 5.1% | ↑ | no event | ✗ |
| May 11 | 6.1% | ↑ | no event | ✗ |
| May 10 | 6.3% | ↑ | no event | ✗ |
| May 9 | 5.9% | ↑ | no event | ✗ |
| May 8 | 5.2% | ↑ | no event | ✗ |
| May 7 | 6.8% | ↑ | no event | ✗ |
| May 6 | 4.9% | — | no event | ✓ |
| May 5 | 4.6% | — | no event | ✓ |
| May 4 | 6.7% | ↑ | no event | ✗ |
| May 3 | 6.4% | ↑ | no event | ✗ |
| May 2 | 5.5% | ↑ | no event | ✗ |
| May 1 | 5.4% | ↑ | no event | ✗ |
| Apr 30 | 5.9% | ↑ | 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/cn — current 7-day calibrated forecast for China with SHAP drivers
- /cn — China country page (current state, recent incidents)
- /sentinel/backtest — global reliability diagram + per-country comparison table
- Calibration refit writeup