Consider a related question: How would you judge the competence of a doctor if you could observe him treating only a single patient?
What you would not do is judge him by the outcome. Even the best physicians have patients die. And even witchdoctors can have patients recover. Randomness is a fact of life (and death). In the case of a medical doctor, the answer seems clear: Instead of looking at the outcome, you would judge him by the decisions he makes and treatments he prescribes. That is, you would examine whether he followed best practices for the circumstances he faced.
It's interesting to consider how standard lists of 'best Presidents' would change if this were used consistently as the criterion (they would certainly become more complicated).