Evidence before a probability.

Current sample and result

1 verified global-reset date; 2 banked grants excluded; 0 global intervals.

Too few verified global-reset intervals. Banked grants are excluded.

What is computed

Counts, the most recent date in this ledger, and observed intervals when at least two distinct global dates exist. Records are deduplicated by date for cadence. This partial history cannot establish the true longest wait.

Before publishing probabilities

Gather a sufficiently complete history, fix a cutoff, save predictions before outcomes, and compare calibration against a simple baseline. Even then, public signals cannot guarantee an official action.