The design and manufacture of most optical systems revolves around the use of ideal optical surfaces. "Perfect" spheres or flats are optimally combined and toleranced during the design phase, and the manufacturers attempt to get as close as possible to these perfect optical surfaces during fabrication. One reason for this stems from the inherent capabilities of the industry's oldest and most pervasive polishing tool: the full-aperture lap. The shape and motion of these tools naturally produce spherical or flat geometries. More recently, a number of new manufacturing technologies based on sub-aperture polishing tools have become available. Sub-aperture tools enable local, preferential removal: a controlled way to polish more material at some locations and less at others. Magnetorheological Finishing (MRF®) is one such sup-aperture polishing technology, and when combined with an accurate measurement, can offer a precise method for converging to the perfect surface: local removal based directly on measured surface height. This capability, however, can also be leveraged in other, more creative, ways. For example, by replacing the typical surface-error measurement by a transmitted wavefront measurement of an entire low-field optical system, a hitmap can be created for one surface in the system that will perfectly compensate for errors of all the other surfaces. This paper will explore a number of examples where "perfectly bad" surfaces have been exploited in actual optical systems to improve performance, improve manufacturability, or reduce cost. In addition, we will ask the question: if making a "perfectly bad" surface was as easy as making a perfectly good one, would this capability be used more widely by the precision optics industry?