Get Overview Trial Retention

Trial-retention survival curve over days since signup: the share of the matured signup cohort still active at each day, Polymorph (treated) vs the Global-control arm. The treated − control gap is the trial drop-off Polymorph prevented. The curve spans a fixed 14-day horizon. With population_in_range the cohort is the recent matured band (signups [now − max(window_days, 30), now − 14]); otherwise it is the whole matured base back to a fixed floor. Empty when global control was never enabled (no randomized arm to compare). Rate-limited per (org, user); the Overview drives it on mount and on window/population change.