What you have said is correct:
I thought the Additional Polling Engine would query whatever hosts I moved to it, so it would share the load with the primary NPM engine?
In the event that the standard poller is up and running, an additional poller polls all the same information from monitored devices that the standard poller polls. The note about EnergyWise, ESX Server, and wireless devices points out the specific features of Orion NPM polling that are dependent upon components installed on the standard poller. In other words, if the standard poller is down, the additional poller, on its own, will not be able to collect EnergyWise, ESX Server or wireless information.
I would suggest that you contact Technical Support on this, they would know the reasons behind the 'why' I would think.
Ugh. Chalk this one up as a "reading comprehension" failure. I see what it's trying to say.
Thanks for the clarification!