@alisonkiddle@mathstodon.xyz Took a different angle to many of the other answers I see - definitely influenced by having worked in data compression! Started thinking about how much entropy the two options are "worth" - more formally, what is -log2(probability) ? A heads on a regular coin flip is "worth" 1 bit of entropy, so ten heads is worth 10 bits. Meanwhile, four sixes on four dice is "worth" 4 * log2(6) bits. But how do we compare those? Well, I know that sqrt(2) is 1.4[something], and in particular is less than 1.5. So 22.5 = 4 * sqrt(2) 10. Converting back to normal people terms, four sixes being "worth" more entropy means that the outcome is less likely. So 10 heads is slightly more likely than 4 sixes.