Uncertainty

A lack of sufficient knowledge or evidence to make a dependable judgment.

Uncertainty can arise from incomplete observations, ambiguous criteria or limits in a model's experience. A system should be able to represent it rather than force every case into pass or fail. Disagreement can indicate uncertainty, but agreement does not eliminate the possibility of shared error.

In practice

When the final placement is off camera, the appropriate label may be unobservable rather than successful or failed.