Sensitivity comparison of searches for binary black hole coalescences with ground-based gravitational-wave detectors

Searches for gravitational-wave transients from binary black hole coalescences typically rely on one of two approaches: matched filtering with templates and morphology-independent excess power searches. Multiple algorithmic implementations in the analysis of data from the first generation of ground-based gravitational-wave interferometers have used different strategies for the suppression of non-Gaussian noise transients and have targeted different regions of the binary black hole parameter space. In this paper we compare the sensitivity of three such algorithms: matched filtering with full coalescence templates, matched filtering with ringdown templates, and a morphology-independent excess power search. The comparison is performed at a fixed false alarm rate and relies on Monte Carlo simulations of binary black hole coalescences for spinning, nonprecessing systems with a total mass of 25-350M, which covers a portion of the parameter space of stellar mass and intermediate mass black hole binaries. We find that in the mass range of 25-100M, the sensitive distance of the search, marginalized over source parameters, is the best with matched filtering to full waveform templates, which is within 10% of the next most sensitive search of morphology-independent excess power algorithm, at a false alarm rate of 3 events/year. In the mass range of 100-350M, the same comparison favors the morphology-independent excess power search within 20% of matched filtering with ringdown templates. The dependence on mass and spin is also explored.