at_risk_counts for left censored data

[danielsadoc]

Hi Everyone, please, how should I interpret the output of at_risk_counts for left censored data? Is there support for left censored data under at_risk_counts?

More generally, what would be a nice way to validate if my KM survival estimator is good for a dataset with left censored data? It turns out that lifelines.utils.find_best_parametric_model is not being able to find any good fit. Are there examples of additional parametric models considered as input for trying

thanks!