You're mistaken. In fact 0.3 is closer to the next-smaller float than it is to the next-larger one which results. But the result of float(0.1) + float(0.2) is n

You're mistaken. In fact 0.3 is closer to the next-smaller float than it is to the next-larger one which results. But the result of float(0.1) + float(0.2) is not any of those things, but rather halfway between the previous and next floats from 0.3, so it gets rounded up instead of down.