Flock operators can label searches “training” or “test.” Those labels appear in audit logs, and they are the only thing distinguishing these searches from any other plate lookup. The new Training & Test Search Patterns report lists every operator who used those labels and shows how their search history looks across time and across plates.
The recurrence score compares how many days an operator searched to the span of time those searches cover. An operator with 200 search days packed into 210 calendar days scores high — their “training” runs continuously with almost no gaps. An operator with 10 search days spread over three years scores low — they searched rarely and stopped. A high recurrence score does not prove misuse, but it narrows the gap between “training” and operational use.
The target-weighted score takes that recurrence and adjusts it for plate diversity. If an operator ran dozens of training searches but hit the same one or two plates each time, the score holds steady. If they searched hundreds of distinct plates, the score rises. Genuine equipment tests tend to use a small fixed set of plates. Searches spread across many different plates look more like real queries wearing a training label.
Filter the report by scope (training only, test only, or both), by state, or by operator name. Sort by either score or by raw active-day count. The report covers all operators in the database who logged at least one search with a training or test label.