False positive for spam

Recently I learned that an e-mail that I sent had been blocked by an over-agressive spam filter. It was a two-line message in which I happened to use the words `finance' and `California', both of which are considered by Bayesian filters to be strong indicators of spam. This is my first personal experience of a communication failure caused by a false positive in a spam filter. Actually, the communication was only delayed in this case, since my correspondent keeps and reviews the filtered messages.