Host Failure Cluster Tolerates & Admission Control

Hi,

I'd like to get some points of view on what seems a common HA configuration - Host Failures Tolerated with Admission Control enabled.

There are no reservations and the slot size is the default.  In a real world scenario where VMs will exceed the resources in a single slot, when would admission control get a chance to work before host resources are exhausted.  In the environment I manage I've used the % resource reservation for the cluster as that seems straight forward and with admission control being of some use.  Just curious as to why this is such a popular option.

Thanks