This paper explores and explains ways to reconstruct reliably human actions and assessments that were part of a failure sequence. It aims to help resist the production of a subjective story from the personal perspective of the investigator. Collating scholarship from the social sciences, psychology, human factors and cognitive systems engineering, it lays out steps to reduce investigative biases and enhance the verifiability of a forensic investigator's conclusions. This may help investigators build and integrate a better description of any human contribution to a failure sequence in their analysis. The ultimate aim of reliably reconstructing human actions and assessments without producing a subjective story from the personal perspective of the investigator could well remain elusive, of course. But steps can be taken to reduce biases and enhance verifiability of conclusions.