A range likely containing the true rate/mean at a stated confidence level (e.g., 95%).
After 200 runs, report a 95% CI for any-unique rate rather than a single estimate.
More data shrinks the interval; high variance widens it. Use exact/binomial methods for small n.