What confidence levels are #
When Loupely or Loupely Lens produces a diagnosis, it assigns a confidence level to the result. This level reflects how strongly the captured signals match a known failure pattern. The confidence level appears in the capture file and sometimes in the popup diagnosis output. Understanding what it means helps you decide how much to rely on the diagnosis and when to verify independently.
High confidence #
A high-confidence diagnosis means the captured signals match a known pattern precisely: the right combination of PHP error type, Hook Execution sequence, browser-side behavior, and timing that Loupely’s correlation rules recognize as a specific failure class. The diagnosis names the specific cause and the triage step is actionable with high certainty. When you see a high-confidence result, follow the triage recommendation directly.
Medium confidence #
A medium-confidence diagnosis means the signals are consistent with a known pattern but not a complete match. Some expected signals are present, others are absent or ambiguous. The diagnosis names the most likely cause but flags that the actual cause could be a related but different failure in the same class. The triage recommendation is still useful, but treat it as the most probable path rather than a definitive answer. If the triage step doesn’t resolve the problem, the next most likely cause in the same failure class is the right place to look.
Low confidence #
A low-confidence result means the signals suggest a general failure category but don’t match any specific pattern closely. The diagnosis may name a failure class (plugin conflict, hook failure, Payment Gateway issue) without being able to identify the specific culprit. Low confidence results are still useful: they tell you which layer of the system to investigate. They just can’t point to the specific fix the way a high-confidence result can.
Null results #
A Null Result means no pattern matched. Loupely captured signals, generated a diagnosis, and found nothing in the correlation rules that matches what it saw. This is different from a failed capture. The capture ran successfully. The signals don’t match known failure patterns. See The Null Result: What It Means and What to Do for how to proceed from a Null Result.