estimate survival.

Enter event times and a 1/0 event indicator (1 = event, 0 = censored). SurvivalFit computes the exact Kaplan-Meier product-limit estimator with Greenwood standard errors and the median survival time, handling right-censoring correctly. Add a group label per subject and, for two groups, it runs the Mantel-Cox log-rank test (χ² on 1 df) to ask whether the survival curves differ. The workhorse of clinical trials, epidemiology, and event-history analysis — real, reproducible, no GPU.