What it measures
False Acceptance Rate (FAR) measures how often a biometric system incorrectly accepts a non-match as a match — mistaking one person's face for another's.
Why it's the number security teams watch most
For identity verification specifically, a false acceptance is the more consequential failure mode — it's the error that lets fraud through, versus a false rejection, which only inconveniences a legitimate user. Systems are usually tuned to keep FAR low even at some cost to FRR.
Where the threshold comes in
FAR and FRR both move in response to where a system sets its confidence threshold for a match — which is why that threshold, not the matching algorithm alone, is a key design decision in any verification system.