voidly

ChinaCN

stable (-3.1pp)

30-day day-by-day shutdown-risk trajectory. Each point is a per-horizon XGBoost prediction (h=1..30) with a 90% conformal band derived from absolute LOCO residuals. Peak risk highlighted.

Peak day
26.8%
d18 · 2026-07-24
Floor day
4.4%
d30 · 2026-08-05
30-day delta
-3.1pp
d30 vs d1 probability

30-day trajectory with 90% conformal band

model trajectory_v1
0%25%50%75%d1d7d14d21d30peak d18 27%
mean P(shutdown by day)90% conformal bandpeak day

Drivers (SHAP top 3 from the 7-day model — same features feed all horizons)

day 17.5%
2026-07-07
  • block_rate_roll30_mean-0.005
  • week_of_year-0.005
  • risk_tier-0.004
day 711.1%
2026-07-13
  • block_rate_roll30_mean-0.005
  • week_of_year-0.005
  • risk_tier-0.004
day 1410.0%
2026-07-20
  • block_rate_roll30_mean-0.005
  • week_of_year-0.005
  • risk_tier-0.004
day 304.4%
2026-08-05
  • block_rate_roll30_mean-0.005
  • week_of_year-0.005
  • risk_tier-0.004

Honest caveats

  • Per-horizon XGBoost stacked into a trajectory — not a true joint distribution.
  • Conformal bands are absolute-residual q90 from LOCO predictions on 18 eval countries.
  • Long horizons (h>=14d) have wider effective uncertainty than the conformal q90 suggests; the same lag features feed all horizons.
  • Bands floor at [0,1]; probabilities below 0.05 are at the model's noise floor.
  • SHAP attribution above is from the 7-day point model. The trajectory uses 30 separate per-horizon XGBoost models that share the same feature set; we have not yet computed per-horizon SHAP.

Sample horizons

DayDateP(shutdown)90% interval
d12026-07-077.5%[0%, 51%]
d72026-07-1311.1%[0%, 76%]
d142026-07-2010.0%[0%, 79%]
d302026-08-054.4%[0%, 83%]
Spotlight countries