Yes. If you have enough radiation dose you'd need more than 100% chance of failure. A much better model is that the number of spots where the chip was destroy

Yes. If you have enough radiation dose you'd need more than 100% chance of failure. A much better model is that the number of spots where the chip was destroyed by radiation is a Poisson distribution with lambda linear in the size of the radiation dose. For low probabilities of failure, the probability of failure is approximately lambda, which is linear. For large lambda, the probability of at least one failure approaches one. However this more accurate model, in the case of interest (low probability of failure), linear is the appropriate approximation. And a small child should get the simpler version and not the complexities of Poisson.