voidly
CN · per-country backtest

China 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

0.000.250.500.751.00threshold 0.50Apr 17Sep 4

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
141
since Apr 17
Accuracy @ 0.5
61.0%
86/141 correct
Brier score
0.005
lower is better
Observed positive rate
0%
mean predicted 6%

All 141 predictions (newest first)

Eval dateForecastPred ≥ 0.5?Observed?Correct?
Sep 42.2%no event
Sep 32.1%no event
Sep 23.1%no event
Sep 12.3%no event
Aug 313.0%no event
Aug 303.6%no event
Aug 291.4%no event
Aug 281.8%no event
Aug 276.8%no event
Aug 2613.5%no event
Aug 259.5%no event
Aug 2414.0%no event
Aug 2320.3%no event
Aug 2223.8%no event
Aug 2111.7%no event
Aug 207.7%no event
Aug 196.8%no event
Aug 186.9%no event
Aug 174.8%no event
Aug 163.6%no event
Aug 1510.8%no event
Aug 143.4%no event
Aug 136.3%no event
Aug 126.6%no event
Aug 116.6%no event
Aug 106.6%no event
Aug 99.8%no event
Aug 810.3%no event
Aug 72.4%no event
Aug 64.9%no event
Aug 54.3%no event
Aug 47.2%no event
Aug 32.1%no event
Aug 25.2%no event
Aug 11.5%no event
Jul 311.3%no event
Jul 301.3%no event
Jul 292.7%no event
Jul 281.2%no event
Jul 276.0%no event
Jul 268.4%no event
Jul 258.7%no event
Jul 244.5%no event
Jul 231.1%no event
Jul 221.5%no event
Jul 212.6%no event
Jul 203.2%no event
Jul 191.9%no event
Jul 181.6%no event
Jul 171.3%no event
Jul 162.8%no event
Jul 152.8%no event
Jul 142.0%no event
Jul 134.3%no event
Jul 1210.1%no event
Jul 1110.2%no event
Jul 108.6%no event
Jul 98.9%no event
Jul 813.9%no event
Jul 75.7%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.
About this view

Every Voidly Sentinel shutdown-risk forecast for China, plotted against what actually happened. 141 (predicted, observed) pairs from the rolling 30-day evaluation window.