voidly
EG · per-country backtest

Egypt forecast vs reality

Every Voidly Sentinel shutdown-risk forecast for Egypt, 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
50.9%
57/112 correct
Brier score
0.433
lower is better
Observed positive rate
64%
mean predicted 41%

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.

0.000.250.500.751.00threshold 0.50Apr 17Aug 6

Y axis: forecast probability · Faint tick: distance to observed outcome (0 or 1)

All 112 predictions (newest first)

Eval dateForecastPred ≥ 0.5?Observed?Correct?
Aug 679.2%shutdown
Aug 578.2%shutdown
Aug 485.8%shutdown
Aug 373.1%no event
Aug 279.4%no event
Aug 168.6%no event
Jul 3160.1%no event
Jul 3058.8%no event
Jul 2957.9%no event
Jul 2856.5%no event
Jul 2756.0%no event
Jul 2647.0%no event
Jul 2541.2%no event
Jul 2448.5%no event
Jul 2352.0%no event
Jul 2252.5%shutdown
Jul 2146.7%shutdown
Jul 2055.5%shutdown
Jul 1959.7%shutdown
Jul 1863.5%no event
Jul 1747.2%no event
Jul 1645.4%no event
Jul 1549.2%no event
Jul 1458.7%no event
Jul 1335.3%no event
Jul 1236.8%no event
Jul 1140.4%no event
Jul 1063.1%no event
Jul 943.5%no event
Jul 864.9%no event
Jul 772.2%shutdown
Jul 663.6%shutdown
Jul 566.8%shutdown
Jul 466.4%shutdown
Jul 384.9%shutdown
Jul 283.4%no event
Jul 178.8%no event
Jun 3069.7%shutdown
Jun 2979.4%shutdown
Jun 2874.0%shutdown
Jun 2775.0%shutdown
Jun 2677.8%shutdown
Jun 2568.4%no event
Jun 2473.6%shutdown
Jun 2369.2%shutdown
Jun 2273.9%shutdown
Jun 2144.4%shutdown
Jun 2043.9%no event
Jun 1943.2%no event
Jun 1844.8%no event
Jun 1745.0%no event
Jun 1643.4%no event
Jun 1543.7%shutdown
Jun 1445.2%shutdown
Jun 1345.4%shutdown
Jun 1244.6%shutdown
Jun 1144.3%shutdown
Jun 1045.0%shutdown
Jun 944.3%shutdown
Jun 844.0%shutdown

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