Azure Autoscaling Application Block - What's the point?

Question

I'm looking at using the Azure Autoscaling Application Block, (or the site management equivalent) - it makes sense on a cloud system to scale according to demand... but there is one thing I don't understand.

The FAQ at the bottom of the Cloud Service Pricing Page warns that

If my deployed instance is in the “stopped” state, do I still get billed?

Yes, a stopped instance still occupies the compute resource that powers the instance,
and therefore you still get billed. 
In order to stop billing, you need to delete your deployment.

So then, what benefit is there in implementing Autoscaling? I have read documentation and blogs... Am I completely misunderstanding something?

Answer 1

At least for Cloud Services (PaaS offering from Windows Azure), when auto scaling kicks in it does not just turns off the deployed instances (Stopped state in Windows Azure terminology), it completely removes it from the pool. Since the VMs are completely removed from your application pool, you only pay for the instances which are running.

So proper utilization of auto scaling could result in not only better resource utilization but also results in cost savings.