Many studies, conducted by different researchers, have shown that in a carefully controlled engine almost all of the cyclic variability, observed in one cylinder, can be attributed to the flame kernel initiation and evolution. The authors have shown that variations in the flame kernel effective area can successfully explain a significant portion of observed cyclic variability. In the search of causes of this variation in effective area, the authors simulated the flame wrinkling in a 2D kinematically simulated turbulent field. That study had shown that flame wrinkling can explain between 5% - 53% of observed variability, depending on the operating conditions. This study explores the effect of adding the third dimension, which in turn allows the addition of flame stretch and curvature effects. Eight of the torque-speed operating points of the previous study are simulated using G-Equation and a 3D kinematically simulated turbulence. The results show that while the geometries of 2D and 3D flames look different the statistics behave quite similarly. The 3D flame shows slightly less variability and can only explain between 3% - 41% of observed variability. Studying the images suggest that this can be attributed to smoothing effects of flame propagation, which is enhanced by incorporating the third dimension. Similar correlation coefficients between flame wrinkling and laminar flame speed and turbulence intensity suggest that the conclusions made for a 2D flame will still be valid for the 3D flame.