The purpose of this in vitro study was to evaluate the impact of the remaining tooth structure on the retention of overdenture cast metal copings. A freshly extracted intact mandibular human canine (length 25 mm) was selected and endodontically treated. An incisal reduction of 4 mm with no ferrule preparation was performed and a post space of 12 mm was created. By using polyvinyl siloxane duplication material and autopolymerizing polymethylmethacrylate resin, ten resin teeth analogs (control group) were obtained. A second set of ten resin teeth analogs (group 1) was created by preparing on the original natural tooth a 360o ferrule design of 1 mm in height and by using the same procedural technique. The canine was further reduced by an additional 1 mm, resulting in a ferrule of 2 mm in height, measured from the initial incisal reduction, thus a third set of ten resin teeth analogs (group 2) was created. For every tooth analog in all groups a dome-shaped metal coping was cast and luted with a glass ionomer cement. All specimens were subjected to tensile load testing until decementation occurred. The specimens in the control group exhibited a mean failure load of 87.21 ± 18.26 N, while the 1 mm ferrule group recorded a higher mean failure load of 125.43 ± 8.79 N and the 2 mm ferrule group recorded the highest mean failure load of 146.12 ± 23.38 N. One-way ANOVA revealed significant differences (F= 28.04, p<0.001) in the tensile failure loads between all of the groups being tested. The metal copings with a 2 mm ferrule design exhibited the highest retention values, followed by the 1 mm ferrule and the no ferrule design, with the differences among them being statistically significant.