[mitthrowaway2]

What we are learning from this thread is that OKR completion percentage is a very bad surrogate metric for OKR difficulty. That's not surprising; it's very hard for managers to evaluate the ambitiousness of individual OKRs unless they are technical experts, whereas completion percentage is an easy surrogate that should correlate with effort and achievement difficulty. Completion percentage is easily and directly measurable, so of course the CEO rewards that instead. Except that it correlates positively with effort and negatively with difficulty, so it cannot simultaneously incentivize working hard on difficult tasks, resulting in terrible counter-incentives and gaming.