Any application that has memory leaks whether active or passive using location services and other system wide services w/o the need of your direct interaction drains the life of the battery charge, period. Craig didn't address any of this in his email and as a former colleague of his I'd love to debate Thermodynamics with him, seeing as he has no background in it.
The debate should begin with respect to the obsession with thin and the fact the battery capacities in iOS devices aren't extending themselves in any impressive ways. They've targeted an 8-9 hour single charge window and expect people [studies proving peaks of use per day drop well before this length] to recharge after.
If they produced a battery of 12 hours it would be impressive, but it wouldn't stop the people who want more from complaining ad nauseum. It's a move the goal post fruitless obsession.
However, so is the obsession with thin. Form factors are done. The system has reached a zenith in design and now it's about incremental refinements from here on out. It's now all about internals.
He assumed normal operating conditions and under such it makes no difference for battery life to kill apps. If some app or apps have leaks or what not you will notice very fast.