What is statistical tolerance analysis (RSS method) and when is it preferred over worst-case?

Statistical tolerance analysis uses the Root Sum Square (RSS) method, assuming tolerances follow a normal distribution. The total tolerance equals the square root of the sum of squared individual tolerances. RSS is preferred when: assemblies have many components (typically 4+), high volume production allows statistical sampling, 100% interchangeability is not required (typically accepting 99.73% yield), and cost reduction from looser tolerances is significant. It provides more realistic predictions than worst-case.