NWA's software supports using "Pattern Rules" on control charts. While control limit violations are triggered when a single subgroup is beyond a control limit, pattern rules are triggered by an unusual pattern of consecutive subgroups. Because patterns involve more than one subgroup, they are more sensitive to smaller, sustained changes than control limit violations.

While the default rules (Western Electric pattern rules) have proven to be effective, one might be tempted to go beyond them and select all of the available rules, or invent new rules. After all, if a few rules are good, even more must be better. While this can be done with NWA software, there is a serious downside.

Adding rules increases the false positive rate. If the pattern rules falsely indicate problems too often, soon they'll be ignored altogether. It's like the boy who cried "Wolf!".

Erroneous control limit violations occur infrequently. Only about three out of a thousand control limit violations are false positives. When a subgroup violates a limit, there's usually a problem.

The Western Electric pattern rules have a much higher false positive rate. About one out of one hundred subgroups will erroneously trigger a pattern rule. So pattern rule violations should be treated more cautiously.

Going beyond the Western Electric rules is hazardous. It's not difficult to add rules so half of the subgroups violate a rule. Authorities on Quality Control techniques recommend removing one or more of the Western Electric rules (often the "4 out of 5" rule), not adding more.

To read more on pattern rules, see: Walker, Esteban and Philpot, John W. and Clement, James (1991) "False Signal Rates for the Shewhart Control Chart with Supplementary Runs Tests" Journal of Quality Technology 23, pp. 247-252.