Experiment exposure records
Record stable assignment, guarded exposure, stopping and cross-period scope.
Prepare an exposure bucket map without gaps or overlap
Assign supplied integer scores to A, B or holdout with explicit boundaries.
Preserve an exposed unit’s variant after an allocation change
Use recorded exposure for existing units and the new map only for unexposed entrants.
Guard experiment start with coherent assignment, render and event tags
Test the complete exposure tuple so rendered B is not recorded as A.
Stop new exposure while retaining the last unit’s observation window
Separate exposure closure from maturity of supplied conversion windows.
Keep repeat participation in separate exposure episodes
Use period-specific exposure identity without treating the same participant as a new person.