The appeal to redefine statistical significance in Nature Human Behaviour reads as a first, reluctant admission by Benjamin and colleagues that there is something wrong with null-hypothesis significance testing. It is of course hard to abandon a practice you have been trained, I would even say conditioned, to use and have been using routinely throughout your professional career. Still, I would argue (see this commentary) that you are being misled, almost as routinely, by the high numbers of false positives, even after the reduction of significance level, and of false negatives, increasing even more after this reduction. There is no context left wherein this erratic practice can be useful for experimental psychology. We should try to stop the theoretical confusion caused by significance testing and focus on more constructive ways to advance psychology.