In this study, we devised a new method to predict facial selectivity by quantifying steric and orbital factors for the nucleophile approaching both π-plane faces. Using this method, we quantified the total electron density and frontier orbital distributions of 163 cyclic ketones with various structures and quantitatively explained the surface selectivity of 323 reactions with eight nucleophiles (BH3, LiAlH4, NaBH4, LiAl(OMe)3H, MeLi, MeMgI, PhLi, and PnMgI). Importance analysis showed a large orbital effect for BH3, LiAlH4, and NaBH4 and the dominance of the steric effect for LiAl(OMe)3H, MeLi, MeMgI, PhLi, and PhMgI. Our method analyzes three-dimensional features based on Gaussian cube files, which can be easily obtained using mainstream computational chemistry software packages, and this approach should prove useful for predicting the rates and facial selectivity of other reactions.