I don't see movement in a hexagon-based grid approaching reality, either. If that's the argument for hexes and against squares, then it's a poor one to make.
In a hex grid, if I move SE then SW, I end up in a very different position relative to my starting point than if I move SE then NE. In fact, the relative difference from the point of origin is large... using the hexagon-grid-generating web site that I linked above, it is a ratio of 168 : 289 pixels, with the first number being how far we ended up to the south by moving SE then SW, and the second number being how far we ended up to the east by moving SE then NE.
Maybe I'm measuring the wrong values here, perhaps because I am using "square-based logic" on a hexagon-based grid, but there appears to be a definite bias against moving east and west.
But, as long as you aren't trying to argue that either model does a good job of following reality, then both have their place and, if for no other reason, it was neat of the Civ 5 team to "change things up" just for the sake of changing things. Not neat in the sense that it makes for a good game, but neat in the sense that it brought in a lot of customers into the Civ series, who would then turn around and start playing Civ 4.