NOTE: I am not a physicist. > They did calculations showing it was impossible Doesn't mean they couldn't have been surprised by something their calculations

NOTE: I am not a physicist. > They did calculations showing it was impossible Doesn't mean they couldn't have been surprised by something their calculations didn't take into account. The yield for the Castle Bravo test in the 1950s was 2.5 times what was predicted [0], because some things they didn't take into account turned out to be important (namely that some of the material in the construction of the device would serve as additional fuel to the explosion). For the first nuclear explosion in 1945, they were initiating an incredibly powerful chain reaction for the first time on Earth. To some extent you just gotta cross your fingers and hope that the chain reaction stops according to your predictions, because you don't know what the "unknown unknowns" are until you try it. The combination of "chain reaction" and "unknown unknowns" is frightening. [0] https://en.wikipedia.org/wiki/Castle_Bravo https://en.wikipedia.org/wiki/Castle_Bravo