In war a soldier behaving properly should take into account a universal requirement not to kill, to be applied strictly in dealing with civilians, but at the same time to support the exception of taking the life of enemy combatants as an act of selfdefense. This is the usual way to distinguish morally the proper treatment to soldiers and civilians. In this article the author criticizes it and outlines a different way to understand and justify the moral distinction mentioned.

Over the past 20 years, there has been a huge amount of work on which things you can kill in self‐defense and why. This paper surveys that work.

Two conditions guiding permissible use of force in self-defence are proportionality and success. According to the proportionality condition the means used to prevent an attack can be permissible only if they are proportional to the interest at stake.1 According to the success condition, otherwise impermissible acts can be justified under the right to self-defence only if they are likely to succeed in preventing the perceived threat.2 These requirements should not always be interpreted narrowly. Sometimes people are permitted to kill culpable (...) aggressors in order to avoid a non-lethal harm. Sometimes people are permitted to take up arms to defend themselves against culpable aggressors they have little or no hope of .. (shrink)

US law continues its historical trend of growing more permissive towards actors who engage in violent action in purported self-defense. We draw on some informal game theory to show why this is strategically irrational and suggest rolling back self-defense doctrines like stand your ground to earlier historical precedents like duty to retreat.