This is utter crap. Someone made a serious error. Including the 42 "heritage" cores in the cost calcuation is hugely distorting; to call it "misleading" is beyond charitable. Someone appers to have been counting--and worse averaging--since the inception of the EELV program (or close to it) for stuff long ago bought and launched.
There are a few of those "heritage" cores under acquisitions prior to the FY2013 block buy still to launch; there are nowhere near 42 (there might be 4). Not only is there no budget for purchase of those hypothetical 42 cores, there are nowhere near enough payloads to make use of them without budget commitments well beyond 2020 (which there are not).
From the USAF FY2016 budget request, the ELC cost through FY2016-2020 is projected to be ~$3B. ELC pays and provides for 8 launches/year (or did), not "approximately 10". NRO pays an additional 25%, so the total for USAF+NRO is ~$4B.
Present and future EELV $400M unit launch cost is much closer to the truth than $225M.
The 42 heritage cores might not be the correct number, but a certain number of heritage cores still will be needed in the equation. You see, the first cores from the infamous block-buy are scheduled for delivery in late 2017.
In between the public announcement of the block-buy contract signing (early 2014) and launch of the first cores from the block-buy is a period of approximately four (4) years. With an average launch rate of 8 missions per year (for USAF and NRO) that adds up to at least 32 cores (not counting any additional cores due to the Delta IV-H configuration).
Even if we use today's date as starting point there will still be the need for at least 24 (and possibly more) heritage cores to cover USAF and NRO launch needs until the first cores of the block buy are put into action:
Having a look over at Gunter's space page I see the need for at least 10 Delta IV cores up to late 2017 and no less than 14 Atlas 5 cores up to late 2017.
So, assuming that the 42 "heritage" cores don't belong in the equation is (at least partially) wrong IMO.