Advanced gravitational-wave observatories, such as LIGO and Virgo, will detect hundreds of gravitational-wave signals emitted by binary black holes in the next few years. The collection of detected sources is expected to have certain properties. It is expected that a selection bias will exist toward higher-mass systems, that most events will be oriented with their angular momentum pointing to or away from Earth, and that quiet events will be much more numerous than loud events. In this paper, we show how all these assumptions are only true for existing detectors and do not have any universality. Using a network of proposed third-generation gravitational-wave detectors, we show how each of these assumptions must be revised, and we discuss several consequences on the characterization of the sources.

• Received 20 October 2016

DOI:https://doi.org/10.1103/PhysRevD.94.121501