You ignore fat-tail risks. The incident in Fukushima is already hugely expensive and the expense keeps growing.
Relying on past frequencies for future risk assessment is dangerous, especially when there is this high of a cost. Especially on the organization level of the decision makers: Maybe on average the risk for the global population is acceptable, but if a country like Germany loses part of its territory to a nuclear accident, that would be huge tragedy for the country and its economy. Germany is NOT replacing nuclear with lignite 1:1, much less in the long term. Anyone who says so is lying. And even those plants MAY have more risk, but that risk is hugely more predictable!
Relying on past frequencies for future risk assessment is dangerous, especially when there is this high of a cost. Especially on the organization level of the decision makers: Maybe on average the risk for the global population is acceptable, but if a country like Germany loses part of its territory to a nuclear accident, that would be huge tragedy for the country and its economy. Germany is NOT replacing nuclear with lignite 1:1, much less in the long term. Anyone who says so is lying. And even those plants MAY have more risk, but that risk is hugely more predictable!