This is the internet so take this with a grain of salt. I work in a life safety field and the last 50 years of research has been to shift the evaluation of safety from an arbitrary value for safety to a statistical basis - all in the name of efficiency (cost, material, environmental - pick you’re cause, they all have a voice). There is generally no perfectly safe condition, only a poly at which the number of standard deviations from the norm makes failure so unlikely as to be nearly impossible. We have classes of prevention and imposed conditions and there are under the intersection of (failure in prevention)x(exceptional imposed danger) are dead people, or at least loss of property.

Shifting the freezer set point is moving the prevention curve. The number may be small, but it’s still finite. The question the actuaries will ask is if the economic value of 19MT is worth an increase in probability for the sickness or death of X people. I’m merely arguing that the value, to me personally, is not sufficient.