Average Run Length (ARL)

Introduction: Average Run Length (ARL)

Average Run Length (ARL) is a statistical measure used in Statistical Process Control (SPC). It represents the average number of observations between two “out-of-control” signals on a control chart. ARL is essential for evaluating the sensitivity and reliability of control charts in detecting process changes.

Background

Control charts, developed by Walter A. Shewhart in the 1920s, are a cornerstone of SPC and quality management. They help monitor process stability by distinguishing between normal variation and unusual patterns. ARL provides a quantitative way to evaluate how effectively these charts detect shifts while avoiding false alarms.

Key Elements / Features

• Definition: Average number of points plotted before a signal occurs.
• High ARL: Indicates a stable process with few false alarms.
• Low ARL: Suggests frequent signals, which may mean high sensitivity or excessive false alarms.
• Balance: Effective SPC requires tuning ARL to be sensitive to real shifts but not overreact to normal variation.

Calculation Formulas:

• Practical (empirical):
  ARL = Total Number of Observations / Number of Out-of-Control Signals
• Theoretical (probability-based):
  ARL = 1 / p
  where p is the probability that any single point signals out-of-control.

Applications / Examples

• Quality Control: Used to assess whether processes remain in control in Lean and Six Sigma environments.
• Process Optimisation: Helps refine control chart parameters to ensure effective detection of process shifts.
• Manufacturing Example: If a production process generates 1 out-of-control signal after every 200 samples on average, then ARL = 200.
• 3-Sigma Chart Example: With a false alarm probability of p = 0.0027, ARL ≈ 1 ÷ 0.0027 ≈ 370.

Relevance / Impact

ARL is a critical performance measure for SPC systems. By monitoring and optimising ARL, organisations can:

• Reduce waste from false alarms.
• Detect genuine process issues faster.
• Improve product quality and process reliability.
• Support continuous improvement in both manufacturing and services.

See also

• Control Charts
• Statistical Process Control (SPC)
• Six Sigma
• Process Capability (Cp, Cpk)