voidly
ER · per-country backtest

ER forecast vs reality

Every Voidly Sentinel shutdown-risk forecast for ER, plotted against what actually happened. 56 (predicted, observed) pairs from the rolling 30-day evaluation window.

Updated every 30 min · CC BY 4.0 · Raw JSON · Current forecast →

Forecasts evaluated
56
since Apr 17
Accuracy @ 0.5
76.8%
43/56 correct
Brier score
0.026
lower is better
Observed positive rate
0%
mean predicted 8%

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 17Jun 8

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

All 56 predictions (newest first)

Eval dateForecastPred ≥ 0.5?Observed?Correct?
Jun 83.5%—no event✓
Jun 75.4%↑no event✗
Jun 63.8%—no event✓
Jun 52.9%—no event✓
Jun 44.9%—no event✓
Jun 34.3%—no event✓
Jun 24.5%—no event✓
Jun 14.9%—no event✓
May 314.4%—no event✓
May 305.1%↑no event✗
May 292.8%—no event✓
May 2880.0%↑no event✗
May 284.7%—no event✓
May 2760.0%↑no event✗
May 273.1%—no event✓
May 264.9%—no event✓
May 255.0%↑no event✗
May 245.7%↑no event✗
May 2360.0%↑no event✗
May 233.6%—no event✓
May 224.7%—no event✓
May 214.9%—no event✓
May 206.1%↑no event✗
May 195.8%↑no event✗
May 184.9%—no event✓
May 175.3%↑no event✗
May 164.2%—no event✓
May 153.8%—no event✓
May 144.6%—no event✓
May 135.4%↑no event✗
May 124.9%—no event✓
May 114.9%—no event✓
May 104.9%—no event✓
May 92.5%—no event✓
May 84.4%—no event✓
May 72.7%—no event✓
May 64.8%—no event✓
May 54.0%—no event✓
May 43.6%—no event✓
May 33.9%—no event✓
May 23.9%—no event✓
May 14.9%—no event✓
Apr 303.3%—no event✓
Apr 293.7%—no event✓
Apr 284.3%—no event✓
Apr 275.2%↑no event✗
Apr 264.0%—no event✓
Apr 253.7%—no event✓
Apr 244.7%—no event✓
Apr 235.4%↑no event✗
Apr 223.9%—no event✓
Apr 214.9%—no event✓
Apr 203.8%—no event✓
Apr 194.5%—no event✓
Apr 184.0%—no event✓
Apr 173.8%—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