I have zero evidence that it's recommended or out performs your scenario. I do have one data point of it working well for me, so I'm planning on doing it again.