GainSeeker analyzes data based on the setting you have selected for Statistical Outliers.
The default setting for this option is set in the Administration module. For more information, see SPC Statistics.
You can temporarily change these settings while generating charts and reports or performing data entry.
If All data included is selected, then all data matching the standard, date range, count, filter, and other retrieval settings you specified—except for data with a Bypass setting of Include on Charts or data lists only or Bypass record—is used to calculate statistics.
Initially retrieving the data
If Data outside Control Limits excluded is selected, then GainSeeker first uses the standard, date range, count, filter, and other retrieval settings you specified to retrieve a set of data. Data with a Bypass setting of Bypass record is not included in this retrieval.
All data retrieved will be displayed in some format on control charts and data tables.
This set of data is reported by the "Total subgroups" and "Total pieces" statistics.
Determining statistical outliers (out-of-control)
From the initial set of data retrieved, GainSeeker then determines which data points are statistical outliers, based on the following criteria for statistical outliers:
Points with a Bypass setting of Include on Charts or data lists only.
Points outside the control limits for 
charts.
These control limits are reported by the "UCLx sub" and "LCLx sub" statistics.
Points outside the control limits for Range charts.
These control limits are reported by the "UCLr sub" and "LCLr sub" statistics.
Note: If you are creating a EWMA chart, Moving Average chart, Trend chart, or other type of chart that uses a different method to calculate control limits, GainSeeker still uses the control limits for and Range charts to exclude outliers. For information on how these control limits are calculated, see Calculating Control Limits.
After excluding these data points from the statistical calculations, GainSeeker recalculates the control limits and then examines the remaining data points to determine if they violate the corrected control limits. This process is of excluding out-of-control data is executed 10 times or until all remaining data points are within the corrected control limits.
Initially retrieving the data
If Out-of-Control, Runs and Trends excluded is selected, then GainSeeker first uses the standard, date range, count, filter, and other retrieval settings you specified to retrieve a set of data. Data with a Bypass setting of Bypass record is not included in this retrieval.
All data retrieved will be displayed in some format on control charts and data tables.
This set of data is reported by the "Total subgroups" and "Total pieces" statistics.
Determining statistical outliers (out-of-control, runs and trends)
From the initial set of data retrieved, GainSeeker next determines which data points are statistical outliers, based on the following criteria for statistical outliers:
Points with a Bypass setting of Include on Charts or data lists only.
Points outside the control limits for charts.
These control limits are reported by the "UCLx sub" and "LCLx sub" statistics.
Points outside the control limits for Range charts.
These control limits are reported by the "UCLr sub" and "LCLr sub" statistics.
Points in a run above the mean on charts.
Points in a run above the mean on Range charts.
Points in a run below the mean on charts.
Points in a run below the mean on Range charts.
Points in an increasing run on charts.
Points in an increasing run on Range charts.
Points in a decreasing run on charts.
Points in a decreasing run on Range charts.
Note: If you are creating a EWMA chart, Moving Average chart, Trend chart, or other type of chart that uses a different method to calculate control limits, GainSeeker still uses the control limits for and Range charts to exclude outliers. For information on how these control limits are calculated, see Calculating Control Limits.
After excluding these data points from the statistical calculations, GainSeeker recalculates the mean and control limits and then examines the remaining data points to determine if they violate the corrected control limits or exhibit a new run or trend. This process is of excluding out-of-control, run and trend data is executed 10 times or until all remaining data points are within the corrected control limits and do not exhibit runs or trends.
On control charts and similar charts, statistical outliers are plotted as diamond markers on the chart. A workstation can also be configured to use a different symbol for statistical outliers by adding the outlier_marker= setting to the Cms.ini file on the workstation. For more information on using this setting, see outlier_marker=.
If you display outliers on histograms, they are included on the histogram with the color and pattern that you choose. To change this setting, click the Options menu, and then click Lines.
In the Special column of a Data Table or other data listing, these points display "SO" ("SO" represents "Statistical Outlier"). The Outlier Reasons column of the Data Table lists the reasons for excluding data points as statistical outliers.
On a list of statistics, the statistical outliers are reported by the "Excluded subgroups" and "Excluded pieces" statistics. Most other statistics (such as Cpk, Ppk, UCL, and LCL) are based on the remaining data values that are not statistical outliers, which are represented by the "Good pieces" and "Good subgroups" statistics.