This THE article shows a remarkable consistency in REF submissions that contain exactly 14 or 24 full time staff. Why should this be? Because, if one submits more than 14 full time staff, then one also has to submit three (rather than two) impact case studies. Having a poor impact case study in one’s submission is much more damaging than not having one (even two or three) members of staff included — regardless of how good their research is. Not being included in a REF submission is a potentially career-damaging event. Even those who unequivocally agree that research should have an impact, and that impact should be measured, should find depressing the unintended effects of the REF rules.