Patterns in numbers fabricated by humans tend to differ from those that occur “naturally” in genuine counts. Statistical procedures have been developed that are sensitive to those differences (but they cannot perfectly distinguish genuine results from fabricated results).
