1 · RESET HISTORY
Start with previous announcements
The forecast uses dated public reset announcements. Broad resets and resets for a named group are recorded separately, but both count when predicting whether any automatic reset will be announced.
METHOD
How the forecast works
The site estimates whether another public automatic Codex reset will be announced within 24 or 48 hours.
2 · TIMING
Compare the current wait
The calculation considers how long it has been since the latest reset and compares that wait with earlier gaps between announcements.
3 · CURRENT SIGNALS
Add strong public hints
A direct statement from a responsible public source can raise the forecast for the time window it describes. The signal and the base forecast remain visible separately.
4 · EXCLUSIONS
Keep unrelated events out
Personal usage refreshes, purchased or banked resets, outages, reposts and unsupported rumors do not count as new public automatic reset announcements.
5 · LIMITS
Treat the result as an estimate
The public history is small and may be incomplete. The percentages are experimental and can be wrong on a specific day. “30%” means roughly three chances in ten across similar forecasts.
If the latest information cannot be confirmed, the site hides the percentage until a fresh update is available. See Research for evaluation results or today’s reset status.