Identifying systematic errors in a data recording system

The present invention differentiates between systematic and non-systematic conditions by observing a figure of merit over a series of many observation events. In a data storage recording environment, the particular figure of merit used is the number of data segments that must be re-written (due to errors) to a recording medium in order to assure that an entire data set is correctly written. A larger number of re-written segments is indicative of a significant error condition. After each data set is completely and correctly written, the number of re-written segments for the data set is reported as an "event." A running history of the classified events (or the events themselves) is maintained. Then, at a predetermined time, the history is analyzed and a decision made as to whether any observed events meets predetermined criteria for a systematic condition.