In an intentional experiment in peer review, the organizers in a computer science conference discover that half of the papers accepted to the conference would have been rejected if the review process were rerun. (Note: in computer science, conference presentations are the meaningful currency of academic productivity, not journal pubs.)