The best protection against random acceleration is a clear thesis about what problem is worth solving and why.