Resources What's New Support Download Contact Sales: Arthur-Industrial@outlook.com
Home › Resources › Probability basics

Probability basics

Events, the addition and multiplication rules, independence, conditional probability and why they matter for quality.

Probability is the language behind every control limit, confidence level and p-value. A probability is a number from 0 (impossible) to 1 (certain).

Rules

  • Complement: .
  • Addition: . For mutually exclusive events, the last term is 0.
  • Multiplication: . For independent events, .
  • Conditional probability: P(A | B) = P(A and B) / P(B).

Quality examples

Rolled throughput yield. A part goes through four steps with first pass yields of 98%, 95%, 99% and 97%. If the steps are independent, the chance a part passes all four the first time is , or 89.4%.

False alarms. A point from a stable, normal process falls outside 3-sigma limits with probability 0.0027. Over 100 points, the chance of at least one false alarm is .

Inspection. If inspection catches 90% of defectives and the process makes 2% defectives, 0.2% of all parts are defective and still shipped (0.02 × 0.10).

Bayes' theorem

. It explains why a test with a low false positive rate can still flag mostly good parts when real defects are rare.

Questions and answers

Are consecutive parts from a machine independent?

Often not. Tool wear, temperature and material lots make neighboring parts similar (autocorrelation). This is why rational subgrouping and checking for autocorrelation matter.

Why does adding more control chart rules increase false alarms?

Each rule adds another way for a stable process to signal by chance. With the four Western Electric rules together, a stable process gives a false alarm about every 92 points on average, instead of about every 370 points with the 3-sigma rule alone.

← All articles