Despite their maturity and popularity, security remains a critical concern in container adoption. To address this concern, secure container runtimes have emerged, offering superior guest isolation, as well as host protection, via system call policing through the surrogate kernel layer. Whether or not an adversary can bypass this protection de- pends on the effectiveness of the system call policy being enforced by the container runtime. In this work, we propose a novel method to quantify this container system call exposure. Our technique combines the analy- sis of a large number of exploit codes with comprehensive experiments designed to uncover the syscall pass-through behaviors of container run- times. Our exploit code analysis uses information retrieval techniques to rank system calls by their risk weights. Our study shows that secure container runtimes are about 4.2 to 7.5 times more secure than others, using our novel quantification metric. We additionally uncover changing security trends across a 4.5 year version history of the container runtimes.