I picked up The Theory That Would Not Die by Sharon Bertsch McGrayne on a whim, a rare dead-tree impulse purchase. And I'm awfully glad I did.
Bayes's Theorem is something you tinker around in S1 - it's the stuff with the $|$ symbol in, about absorbing information you already know into your probability calculation. But, surprisingly, Bayes' Theorem isn't all that old (it's one of the few mathematical things to have come out of Britain's post-Newton slump) and even then it was pretty much abandoned by Bayes himself. It was Laplace who did all the work, and even he was laughed out of town.
It really wasn't until World War II, when Turing and Good started applying it to code settings, that Bayes' rule got any sort of positive reception; unfortunately, their work was classified so it took another 30 or 40 years before it started to be used properly - for things like catching submarines, finding nuclear missiles and automated driving.
It's not an extremely mathematical book - far from it, it's much more of a history - although there's a nice appendix with a few examples of Bayes being applied at the back that I found quite instructive. I found it very readable, and at a level an A-level student (or a motivated GCSE student) could enjoy.
I know it's a non-fiction book. I know Bayes' Theorem is not a character in a story. But I still found myself rooting for it as it struggled to show what it could do.