The match for duplicate data functionality is used to generate matches for duplicates of all non-system base object columns. These matches are generated when there are more than a set number of occurrences of complete duplicates on the base object columns. For most data, the optimal value is 2.
Although the matches are generated, the consolidation indicator remains at 4 (unconsolidated) for those records, so that they can be later matched using the standard match rules.
The Match for Duplicate Data job is visible in the Batch Viewer if the threshold is set above 1 and there are no NON_EQUAL match rules defined on the corresponding base object.