In mammals, high dominance rank among males is often associated with mating success. However, mating opportunities do not automatically translate into offspring production; observed mating success may be discordant with offspring production, for three reasons.(1) Observed mating may be nonrepresentative of actual mating if some mating is surreptitious (decreasing the chance that it will be observed), (2) mating may be nonrandom if some males allocate more mating effort to females with high fertility (i.e. if some males differentially invest in higher fertility mating) and (3) conception success may be nonrandom if sperm competition or sperm selection play a role in conception. We performed a genetic analysis of paternity in the well-studied savannah baboon, Papio cynocephalus, population in the Amboseli basin, eastern Africa, in order to measure the concordance between observed mating success and actual offspring production. We found that observed mating success was generally a good predictor of paternity success, that high-ranking males had higher paternity success than lower-ranking males, and that male density and male rank stability contributed to variance in male paternity success. We found little evidence for successful surreptitious mating (although subadult males did occasionally produce offspring, apparently using this strategy), and no clear evidence for differential sperm success or sperm depletion (although we could not rule them out). However, we found clear evidence that high-ranking males showed mate choice, concentrating their mating efforts on females experiencing conceptive rather than nonconceptive cycles.