So this county has voted for the national winner in every election from 1964-2012.
It's safe to assume, then, that they got 1960 wrong, otherwise it would have been included in the sample. This means that they are less than 93% "accurate" over the last 14 elections.
As there are several thousand counties in the US it's surprising not that this one was "nearly perfect", but that there were not several that did better.