The purpose of this study was to compare the rollback ratio in bicruciate substituting (BCS) total knee arthroplasty (TKA) and bicruciate-retaining Oxford unicompartmental knee arthroplasty (UKA). In this study, 64 subjects (64 knees) undergoing BCS-TKA (Journey II: Smith and Nephew) and 50 subjects (50 knees) undergoing Oxford UKA (Zimmer-Biomet holdings, Inc., IN) were included. Approximately 6 months after surgery, and when the subjects had recovered their knee range of motion, following the Laidlow's method, lateral radiographic imaging of the knee was performed with active full knee flexion. The most posterior tibiofemoral contact point was measured for the evaluation of femoral rollback (rollback ratio). Flexion angle was also measured using the same radiograph and the correlation of rollback and flexion angle was analyzed. As a control, radiographs of the asymptomatic contralateral knees of subjects undergoing Oxford UKA were evaluated (50 knees). The rollback ratios of the BCS-TKA, Oxford UKA, and control knees were 37.9 ± 4.9, 35.7 ± 4.2, and 35.3 ± 4.8% respectively. No significant difference in rollback ratio was observed among the three groups. The flexion angles of the BCS-TKA, Oxford UKA, and control knees were 123.8 ± 8.4, 125.4 ± 7.5, and 127 ± 10.3 degrees, respectively. No significant difference in knee flexion angle was observed among the three groups. Significant correlation between rollback ratio and knee flexion angle was observed (p = 0.002; Pearson's correlation coefficient = − 0.384). BCS-TKA showed no significant difference in rollback ratio when compared with control knees and Oxford UKA knees. The BCS-TKA design is likely to reproduce native anterior cruciate ligament and posterior cruciate ligament function, and native knee rollback.