Unless you're making the turntable completely redundant, the axial accuracy is always going to be lower the farther you get from the center (though the maximum print speed increases proportionally). Using a "Delta-style" y and z axis could be used to give you higher radial resolution at the edge, but it doesn't change the axial resolution at all, (likewise, the worse radial accuracy in the center doesn't do anything to change the better axial resolution of the bed).

Basically, though the minimum level of accuracy for either system is close enough to the average accuracy that the difference isn't particularly noticeable.

The big challenge to polar coordinate printers becoming mainstream is that lines passing through the center point of the printer have to either dwell in place there for a significant amount of time or move the plate very fast. A dedicated polar-coordinate slicer can handle this problem easily, and that looks like the direction that Polar 3D took in this case. Open source projects like the R360 reprap and other polar-style RepRaps tried too hard to be compatible with Cartesian-based slicers, doing the conversion math in the firmware and running into major problems when the math comes up with an unreasonable figure, as is the case with lines passing through the center.