Aedes aegypti (L.) is an important vector of viruses causing dengue, Zika, chikungunya, and yellow fever and as such is a threat to public health worldwide. Effective trapping methods are essential for surveillance of both the mosquito species and disease presence. The BG-Sentinel (BGS) is a widely used to trap Ae. aegypti but little is known of its efficiency, i.e., what proportion of the mosquitoes encountering the trap are captured. The first version of the BGS trap was predominantly white, and the current version is mostly navy blue. While this trap is often deployed without any olfactory lure, it can also be deployed with CO₂ and/or a human skin odor mimic lure to increase capture rates. We tested the efficiency of capturing Ae. aegypti under semi-field conditions for the original white version without lures as well the blue version with and without various lure combinations. None of the configurations tested here captured 100% of the mosquitoes that encountered the trap. A navy-blue trap emitting CO₂ and a skin odor mimic produced the highest capture (14% of the total insects in the semi-field cage), but its capture efficiency was just 5% (of mosquitoes encountering the trap). Mosquitoes often had multiple encounters with a trap that did not result in capture; they crossed over the trap entrance without being captured or landed on the sides of the trap. Understanding these behaviors and the factors that induce them has the potential to suggest improvement in trap design and therefore capture efficiency.