We present a deterministic toy forecast model and give an elementary proof of why it is not reliable when used to make long-term predictions.