This work provides a statistical analysis of the massive star binary characteristics in the Cygnus OB2 association using radial velocity information of 114 B3-O5 primary stars and orbital properties for the 24 known binaries. We compare these data to a series of Monte Carlo simulations to infer the intrinsic binary fraction and distributions of mass ratios, periods, and eccentricities. We model the distribution of mass ratio, log-period, and eccentricity as power laws and find best-fitting indices of {alpha}=0.1+/-0.5, {beta}=0.2+/-0.4, and {gamma}=-0.6+/-0.3, respectively. These distributions indicate a preference for massive companions, short periods, and low eccentricities. Our analysis indicates that the binary fraction of the cluster is 44%+/-8% if all binary systems are (artificially) assumed to have P<1000 days; if the power-law period distribution is extrapolated to 10^4^ years, then a plausible upper limit for bound systems, the binary fraction is ~90%+/-10%. Of these binary (or higher order) systems, ~45% will have companions close enough to interact during pre- or post-main-sequence evolution (semi-major axis <~4.7AU). The period distribution for P<26 days is not well reproduced by any single power law owing to an excess of systems with periods around 3-5 days (0.08-0.31AU) and a relative shortage of systems with periods around 7-14 days (0.14-0.62AU). We explore the idea that these longer-period systems evolved to produce the observed excess of short-period systems. The best-fitting binary parameters imply that secondaries generate, on average, ~16% of the V-band light in young massive populations. This means that photometrically based distance measurements for young massive clusters and associations will be systematically low by ~8% (0.16 mag in the distance modulus) if the luminous contributions of unresolved secondaries are not taken into account.

Cone search capability for table J/ApJ/751/4/stars (Stars used in the Monte Carlo analysis (table added by CDS))