Recently, the LIGO, Virgo and KAGRA scientific collaboration reported the detection of nearly 80 gravitational wave sources, with the majority of them being binary black holes. The increased number of binary black hole observations allow for a more detailed investigation of their mass, spin and redshift distributions. I will discuss these distributions and consider how they can (or cannot) be interpreted in terms of current astrophysical models. Specifically, I will consider the dynamical formation of black hole binaries in globular clusters and present recent results from our new fast cluster code, cBHBd. Our models indicate that a purely globular cluster formation scenario for the detected black hole binaries is inconsistent with current data. On the other hand, we find that a cluster origin might explain the most massive mergers detected so far, as well as the lack of a high mass cut–off in the inferred mass distribution.