Sometimes, punctuation in captions can provide important clues about what's going to happen, regardless of how well or poorly timed the captions are.
Which sounds are significant? How does the captioner choose which sounds to caption? Are some captions unnecessary? Why isn't it possible to caption every sound in the environment?
In this example, the caption user recognizes a heartbeat before the non-caption user that because the bad guy's captioned sentence is unfinished ("We can nego-"), he will be shot before he can finish saying "negotiate."