Real fights/self-defense encounters tend toward being chaotic and sloppy. This is true irregardless of the weapons(or none) used. Comparing what you see at the range with what we see in actual gunfights or gun vs knife/clubs/bare hands, reminds me of the differences we see when we compare what is practiced in many so-called "traditional" martial arts schools with the dynamics of actual fights. The traditional martial artist will often watch a real fight and scoff at the lack of technique and training of the participants. They did the same with NHB/UFC/MMA when it first came to the public attention and many still do. Yet they usually always refuse to demonstrate the effectiveness of their method when asked to do so. The problem is the hyper-structured and robotic training and techniques many traditional martial arts advocate is not based in reality nor usually all that effective in actual fights.
The same applies to those whose firearm training is limited to static range work. We simply do not see the controlled, robotic movements and picture perfect stances common in range work with any regularity in Force-on-Force training nor in video footage of actual defense scenarios involving gunfire. A lot of folks have trouble accepting that, but the truth is the truth even if it is dismissed and ignored. A lot of people want to look to what happens in Police work, but aside from a few exceptions, I think it has limited value to the armed civilian since it's a matter of reactive vs proactive, concealed vs open carry, avoid and escape vs intentionally engage. As an armed civilian, I have to assume any scenario I'm involved in will be reactive otherwise I would simply avoid it. I may have to engage at contact distances and be forced to incorporate empty hands skills while accessing or even while using my firearm. The safety would already be off at this point, so would be of no use. Anyone who has attended an ECQ class knows how chaotic these scenarios can get. And anyone who has participated in a lot of FoF also has seen people fail(not just forgetting)to get a manual safety disengaged during reactive scenarios.
I may have to move or flat out run with my weapon drawn, even doing so while trying to corral and direct my family to safety. Depending on the specific situation, my finger may or may not be on the trigger. If I'm investigating a bump in the night at home where I think someone could be lying in wait, my finger will indeed be on the trigger when I check it out. I could trip and fall in such a situation or it could quickly become a contact struggle. If I had a weapon with a safety, it would again be off and of no use. A heavier trigger pull may or may not prevent an unintentional discharge, but it definately would in many of them. In terms of some type of startle-flinch response, the available data says it still will prevent unintentional discharges in many cases.
Force Science News #3: Can You Really Prevent Unintentional Discharges? plus I'm nowhere near convinced that some type of sympathetic reflexive hand clenching is necessarily a widespread hard-wired reflex. For me, no manual safety and a heavier trigger pull logically constitutes the best compromise between readiness and safety.