kbitzer's (awesome) graph is only representing the lower left quadrant of the full graph, because that's where everyone's scores fell. so technically the exact center would be the upper right corner of his graph, making Spindaddy and Bdoomed the most "center" of those graphed. 
But what makes that place the "center"? The person(s) who designed the test called that "center" and put it in the middle of the graph, but why that place? You could draw the graph any way you wanted and each time there'd be a different center. There's nothing objectively "center" about that middle point on the test's graph. That's what I'm saying. No matter what graph you draw, you're forcing a particular center to appear, "center" is not inherent in the actual data, it's an artifact of a particular interpretation.
Okay, now I see better what you were getting at. Based on the website's graph of various political leaders (historical and recent), it looks like they've simply made the various extremes determine the boundaries of the graph and thus the center is defined by the breadth of left and right over the past 100 years or so. This strikes me as a fairly useful rubric because no one period of right- or left-leaning should be enough to shift the entire graph.
However, I'm guessing you might argue that 100 years is still a relatively limited time frame, and I agree that it would be interesting to see if adding the major political leaders of known empires would change the graph at all. For instance, where would Julius Caesar and King John of England (who was forced to sign the Magna Carta) fall on this graph? Which begs the question, do we really know enough about some of the historical leaders to accurately place them, and does the concept of right and left even apply to the politics of their time?
There's nothing objectively "center" about, say, not totally agreeing with Rs or Ds. In, say, a climate where the right-leaning bit of the sample has shifted way, way to the right, "I don't totally agree with either side, I go right down the middle" actually puts one further right than this graph would call "center."
I'm just saying, "center" with no other points of reference (opinion is split between R and D, score on the political compass test, in relation to a good sample of the population of whatever country you vote in, whatever) is meaningless.
Well, as Iamafish has reminded us, just because the center of
our graph is falling to the left of the graph's center, that doesn't mean we are proving the rubric to be at fault. I know plenty of people who would fall squarely in the upper right quadrant, whose scores would serve to balance ours and place the center back in the center of the graph.
