Knee replacement is a proven and reproducible procedure to alleviate pain, re-establish alignment and restore function. However, the quality and completeness to which these goals are achieved is variable. The idea of restoring function by reproducing condylar anatomy and asymmetry has been gaining favor. As knee replacements have evolved, surgeons have created a set of principles for reconstruction, such as using the femoral transepicondylar axis (TEA) in order to place the joint line of the symmetric femoral component parallel to the TEA, and this has been shown to improve kinematics. However, this bony landmark is really a single plane surrogate for independent 3-dimensional medial and lateral femoral condylar geometry, and a difference has been shown to exist between the natural flexion-extension arc and the TEA. The TEA works well as a surrogate, but the idea of potentially replicating normal motion by reproducing the actual condylar geometry and its involved, individual asymmetry has great appeal. Great variability in knee anatomy can be found among various populations, sizes, and genders. Each implant company creates their specific condylar geometry, or “so called” J curves, based on a set of averages measured in a given population. These condylar geometries have traditionally been symmetric, with the individualised spatial placement of the (symmetric) curves achieved through femoral component sizing, angulation, and rotation performed at the time of surgery. There is an inherent compromise in trying to achieve accurate, individual medial and lateral condylar geometry reproduction, while also replicating size and avoiding component overhang with a set implant geometry and limited implant sizes. Even with patient-specific instrumentation using standard over-the-counter implants, the surgeon must input his/her desired endpoints for bone resection, femoral rotation, and sizing as guidelines for compromise. When all is done, and soft tissue imbalance exists, soft tissue release is the final, common compromise. The custom, individually made knee design goals include reproducible mechanical alignment, patient-specific fit and positioning, restoration of articular condylar geometry, and thereby, more normal kinematics. A CT scan allows capture of three-dimensional anatomical bony details of the knee. The individual J curves are first noted and corrected for deformity, after which they are anatomically reproduced using a Computer-Aided Design (CAD) file of the bones in order to maximally cover the bony surfaces and concomitantly avoid implant overhang. No options for modifications are offered to the surgeon, as the goal is anatomic restoration. Given these ideals, to what extent are patients improved? The concept of reproducing bony anatomy is based on the pretext that form will dictate function, such that normal-leaning anatomy will tend towards normal-leaning kinematics. Therefore, we seek to evaluate knee function based on objective assessments of movement or kinematics. In summary, the use of custom knee technology to more closely reproduce an individual patient's anatomy holds great promise in improving the quality and reproducibility of postoperative function. Compromises of fit and rotation are minimised, and implant overhang is potentially eliminated as a source of pain. Early results have shown objective improvements in clinical outcomes. Admittedly, this technology is limited to those patients with mild to moderate deformity at this time, since options like constraint and stems are not available. Yet these are the patients who can most clearly benefit from a higher functional state after reconstruction. Time will reveal if this potential can become a reproducible reality