In my opinion the idea that a bug bash is necessary to find bugs points out a flaw in test coverage or test methodology. The rationale being if test cases are written which provide adequate test coverage and these test cases are executed (through manual or automated test means), there shouldn't be much else left for an ad-hoc bug bash to find.

Automating a bug bash makes no sense to me. Shouldn't that automation simply be regular test automation?