Comment by sbirch

My favorite explanation of this (posed instead as the Locomotive problem) is in Allen Downey's "Think Bayes," pp.22

It's online too, and worth reading!

http://www.greenteapress.com/thinkbayes/