Experimental model to illustrate the problem of gun violence as a 'wicked problem'.
Basic idea: Guns exist. Murders exist. When people become aware of murders, they will tend to buy guns at a higher rate. Problem is, the more guns that are out there, the more likely it is that a gun will be used in a murder. People hear about increasing gun murder rates, they buy more guns. Problem feeds itself.
Now, accidents also happen, at some rate, as an inevitable consequence of gun ownership.
This model inter-relates gun supply as a function of purchases (with some attrition included), with purchases a function of replacing attrition, accidents (depress purchases), and murders (increase purchases).
Accidents tend not to be as visible, so they have a lower impact than murders on purchases. Further complication: the more murders, the greater odds that people hear about them, and buy guns to protect themselves at higher rates.
End result, even if you arbitrarily max out gun ownership at, say, 50% of the population, you've got an effect where murders are far more common than they were. Reversing the effect is not linear, due to the way I've set up the purchases logic: higher purchase rates with higher thresholds of murders (1,5,10+)
Idea of showing this logic: if you want to reduce gun purchases, murder rate needs to come down. But, to accomplish this, you need to get guns out of circulation. Problem: positive effects won't be immediately visible, because the interactions are non-linear.
I think this helps illustrate why the US gun problem is so severe. Inclusion of the accident effects may offer an explanation for differences between, say, Switzerland (high gun ownership) and the US: awareness of accident risk.
This is ENTIRELY ad-hoc, no real parameters were used. For illustration of the theoretical problem ONLY. Though if you can come up with better parameters rooted in the relevant science literature (I may do it someday, but not the time right now...), you might be able to turn this into a more 'real' simulation.
