Increased access to large datasets of real-world drive cycles is driving demand for vehicle powertrain models capable of rapidly estimating real-world fuel economy. Whether for component design tradeoff studies or regulatory analysis, the need for powertrain models to achieve high levels of accuracy with low runtimes is critical. One approach is to develop simplified models that can be calibrated to controlled laboratory testing. However, many of the factors impacting real-world fuel economy are often left unexplored in the controlled laboratory setting. This paper seeks to quantify the ability of a simplified vehicle model, calibrated to laboratory test data, to accurately estimate real-world fuel economy in an uncontrolled, on-road environment. Model validation results from over 2,500 miles of on-road testing are presented for a representative, conventional gasoline, mid-size sedan equipped with laboratory-grade instrumentation.