I'm comparing only candidates which brought voters to the polls with the expectation that their vote might actually make a difference in 2012 vs. 2008.
The raw turnout comparison is counting all candidates, even those who polled less than a quarter million actual votes.
So basically you’re saying that the 2012 turnout exceeds the 2008 turnout if you exclude the 300,000 more people who voted in 2008.