This paper examines the effects of increased car weights, train size, and distance on railroad operational performance, which is measured in revenue ton-miles per train hour. A statistical model is estimated from Class I data that explains most of the variation in this index, while controlling for commodity and fixed network effects. The results suggest that length of haul, net load, and train size contributed significantly to performance gains from 1987 to 2006. However, train miles per track mile increased by 89% during the same period, resulting in line and yard congestion.