In safety-oriented industries, a model that is frequently used is called the "Swiss Cheese Model". The idea is that each layer of safety you incorporate into your process is like a layer of swiss cheese with holes in it and these layers of cheese are stacked upon one another randomly. If all the holes line up, then your safety system (or mechanisms) did not prevent the incident from becoming an accident.
To apply that concept to a handgun, you could add each mechanical safety as a layer of cheese. You could probably also argue that all holes are not created equal. Where as a 2lb SA trigger with 1/8" travel may be represented as a large hole in the slice, a 20lb DA trigger with 1" of travel might be a very small hole in the slice.
So in general, the more slices and the smaller the holes, the safer the design. So something like a 1911, I might give three slices: the light, crisp SA trigger would be one slice with a rather large hole, the thumb safety and grip safeties would each get a slice with a relatively small hole. For something like a Glock, I'd give it two slices: one for the blade on the trigger and one for the trigger pull itself which is really a Light Double Action style that I'd give a smaller hole compared directly to the 1911's trigger pull. Overall though, I'd say the three layers provided by the 1911 would still be safer than the two layers provided by the Glock. Anyway, that's just an illustration of how I personally would model it [and yes I'm aware of all the various internal safeties but didn't include them since things like firing pin blocks are really more for dropping the gun].
Just a thought. The world of HES has come up with many pseudo-scientific ways to model risk, but I've always liked the swiss cheese model as it is easy for people to visualize.
