Cloud signals from a vertically pointing, range-corrected ruby lidar and a rotating beam ceilometer showed excellent agreement in the height at which peak signal occurred. However, pilot reports of ceiling were at significantly lower altitude when viewed at 3° below horizontal without the aid of approach lights. An analytical relationship is derived that connects pilot's optical depth with an idealized lidar's signal profile by applying reasonable approximations to the cloud structure. The view angle of the pilot, the lidar's pointing direction, and the vertical distribution of the obscuring particles control the relationship. This relationship is recommended as the foundation of a signal-processing algorithm for better ceiling measurements, particularly for low stratus. Multiple scattering and a ceilometer's range dependence are complicating factors. Accurate cloud base predictions should also account for horizontal inhomogeneities at the cloud's bottom as shown by simple examples. An of these concepts are also pertinent to measurement of slant visual range.