Instrumented-pointers are often used to calibrate anatomical landmarks in biomechanical analyses. However, little is known about the effect of altering the orientation of the pointer during calibration on the co-ordinates recorded. Incorrect positioning of a landmark influences the axes created, and thus the kinematic data recorded. This study aimed to investigate the reliability of the pointer method for anatomical calibration. Two points were drawn onto a fixed box to resemble knee joint epicondyles, then a custom-made pointer was used to define the positions of these landmarks in three-dimensions. Twenty different pointer-orientations were chosen, and the position of the pointer in each of these orientations was recorded 8 times. Euclidean distances between single points were calculated for both landmarks and compared statistically (α = 0.05). Average Euclidean distances between all reconstructed points were 3.2±1.4mm (range: 0.3-7.1mm) for one landmark and 3.3±1.5mm (range: 0.3-7.9mm) for the other. The x- and y-co-ordinates recorded differed statistically when the pointer was moved about the X and Y axes (anterior/posterior and superior/inferior to landmark) (p 0.05). ICC values for all co-ordinates were excellent, highlighting the reliability of the method (ICC > 0.90). These results support this method of anatomical calibration; however, we recommend that pointers be consistently held in a neutral oriented position (where the handle is not anterior, posterior, superior or inferior to the landmark) during calibration, to reduce the likelihood of calibration errors.