If you can provide statistics on bugs being an order of magnitude higher in dynamic languages to back up your figures here, it would be a far more convincing argument. I'm sure type safety catches a few bugs and prevent some from being made, but does it catch as many as you think?