Looks like you've discovered that KBB applies a linear mileage cost. Does that reflect reality? I doubt it -- at least, not after you reach a certain mileage.
I'm curious if you grabbed data to extend that chart out to, let's say, 200k miles. There should be a sigmoid somewhere out there.
http://www.undiscoveredfeatures.com/2011/09/kelly-blue-book....
What I found interesting was comparing the rate of depreciation per mile vs time.