Understanding capability measures

There are two basic types of capability measures: long-term (LT) and short-term (ST).

LT means that the process has had ample opportunity to exhibit typical shifts and drifts, cyclical patterns, and opportunities for special causes of variation. LT is meant to measure how the process actually performs over a long period of time in the presence of shifts, drifts, cyclical patterns, special causes, and so on.

ST means that the process is devoid of those characteristics. ST is meant to measure the potential capability of the process if the process were performing at its absolute best.

A process has four possible states or conditions, as described in the following sections. When you calculate ST capability statistics or indices, use condition A. When you calculate LT capability statistics and indices, use condition D.

Condition A

Condition A depicts the process performing at its optimum. Here, the process is on target and does not exhibit any type of dynamic drifting or other patterns. Hence, Condition A best portrays the potential capability of the process, because every process output has the potential to be on target and to be devoid of dynamic patterns and drift.

• The process is centered on target.
• The process does not exhibit shift or drift.
• The process does not exhibit dynamic patterns.

Condition B

Condition B does not portray the absolute potential of the process because it is not on target.

• The process is not centered on target. That is, the process shifted.
• The process does not exhibit drift.
• Process does not exhibit dynamic patterns.

Condition C

Condition C is not usually an accurate portrayal of the actual performance of a process over a long period of time, because it assumes that the process is on target and does not account for shifts in the process mean.

• The process is centered on target.
• The process exhibits drift.
• The process exhibits dynamic patterns.

Condition D

Condition D better depicts the actual performance of the process over a long period of time. The process is not assumed to be on target and exhibits dynamic patterns and drift.

• The process is not centered on target.
• The process exhibits drift.
• The process exhibits dynamic patterns.