voidly
SY · per-country backtest

Syria forecast vs reality

Every Voidly Sentinel shutdown-risk forecast for Syria, 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
42.9%
48/112 correct
Brier score
0.330
lower is better
Observed positive rate
14%
mean predicted 33%

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 681.5%no event
Aug 574.3%no event
Aug 474.2%no event
Aug 34.1%no event
Aug 282.3%no event
Aug 179.2%no event
Jul 3181.5%no event
Jul 3085.5%no event
Jul 2989.7%no event
Jul 2877.1%no event
Jul 2775.3%no event
Jul 2677.8%no event
Jul 2580.2%no event
Jul 2487.4%no event
Jul 2365.3%no event
Jul 2272.1%no event
Jul 2173.2%no event
Jul 2069.0%no event
Jul 1970.9%no event
Jul 1865.3%no event
Jul 1771.6%no event
Jul 1663.6%no event
Jul 1572.2%no event
Jul 1470.3%no event
Jul 1365.8%no event
Jul 1270.4%no event
Jul 1166.8%no event
Jul 1061.7%no event
Jul 964.3%no event
Jul 876.2%no event
Jul 775.5%no event
Jul 668.2%no event
Jul 560.0%no event
Jul 466.4%no event
Jul 370.5%no event
Jul 275.7%no event
Jul 181.4%shutdown
Jun 3081.8%shutdown
Jun 2980.1%no event
Jun 2887.9%no event
Jun 2782.2%no event
Jun 2682.2%no event
Jun 2571.7%no event
Jun 2463.3%no event
Jun 2364.9%no event
Jun 2257.5%no event
Jun 212.9%no event
Jun 205.4%no event
Jun 193.8%no event
Jun 183.8%no event
Jun 172.9%no event
Jun 164.6%no event
Jun 154.1%no event
Jun 143.9%no event
Jun 134.8%no event
Jun 124.0%no event
Jun 114.5%no event
Jun 104.7%no event
Jun 94.5%no event
Jun 84.8%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