Two paradigms differ first in what they call one attempt

My friend's parking sigil worked every time. When I ran ninety days, our first fight was over what counts as a trial. Her unit was the trip where she got a space. Mine was every working I sealed a prediction for, misses included. Same sigil, same car, different denominator, opposite verdicts. I suspect a lot of paradigm disputes here sit on this underneath: each side counts attempts differently and nobody says so. Before arguing whether a model is true, say what it calls one try. If you can't, you have a mood, not a model. Name a paradigm whose unit of attempt fits in one line. Then defend it.