This paper developed and compared two machine learning models - the Explainable Boosting Machine (EBM) and Quantile Regression Forests (QRF) - to predict acetabular cup size for total hip arthroplasty (THA). The study utilized patient data from 30,587 THA procedures at a single medical institution (from March 2016 and January 2024). Both models were trained on nine easily accessible preoperative variables, such as height, sex, age, weight, race, anterior or posterior surgical approach, computer navigation, and robotic assistance. QRF demonstrated superior performance, correctly predicting 82.85% of cups within plus or minus 2 mm and 97.27% of cups within plus or minus 4 mm. Sex, height, age, weight, surgical approach, and BMI were deemed the most influential factors in the QRF model. Partial dependency plots were also generated from the EBM model, analysis of which showed that male sex, greater height and age, posterior surgical approach, and lower BMI were associated with a larger predicted cup size. The results address key limitations of previous models by providing high accuracy and prediction using convenient patient information.