On fairness of completion-based theorem proving strategies