Skip to main content

Process Capability (Cpk/Ppk)

Process capability indices (Cpk and Ppk) quantify how well a manufacturing process can produce parts within specification limits. Cpk measures short-term capability using within-subgroup variation, while Ppk measures long-term performance using overall variation.

Why It Matters

Process capability is the universal language between quality engineers, customers, and suppliers. An automotive OEM requiring Cpk ≥ 1.33 is saying: "Prove your process produces fewer than 63 defects per million opportunities."

Without reliable capability indices, suppliers cannot bid on contracts, production teams cannot validate process changes, and quality managers cannot prioritize improvement efforts. Cpk/Ppk drives everything from PPAP submissions to annual quality reviews.

The catch is that traditional Cpk calculation assumes your data is normally distributed. When data is skewed, heavy-tailed, or multimodal — which is common in real manufacturing — the standard formula produces misleading indices that overstate or understate true capability.

The EntropyStat Perspective

EntropyStat computes capability indices without assuming normality. Instead of forcing data into a bell curve and deriving Cpk from the mean and standard deviation, the EGDF (Entropic Global Distribution Function) learns the actual shape of your process distribution directly from measurements.

This matters most with small production runs. Traditional Cpk requires 30+ measurements to produce a stable estimate because the normal distribution parameters (mean, sigma) converge slowly. EntropyStat's entropy-based approach produces reliable capability estimates with as few as 5–8 measurements because it does not need to estimate parametric distribution parameters at all.

When the ELDF (Entropic Local Distribution Function) detects hidden clusters — for example, parts from two slightly different tool inserts — EntropyStat reports separate capability indices for each subpopulation. Traditional Cpk would average them together, masking a potential out-of-spec cluster behind an "acceptable" aggregate number.

Try our free Cpk calculator → to compute traditional Cpk, Ppk, and Cp instantly — then see what entropy-based analysis reveals about your data.

Related Terms

See Entropy-Powered Analysis in Action

Upload your data and compare traditional SPC with entropy-based methods. Free demo — no credit card required.