The Weather Research and Forecast (WRF) model has been designed for use with grid scales from 1 to 10 km and to be the next-generation operational and research model. WRF's highly-touted success in forecasting Isabel, suggests WRF may help improve tropical storm forecasts. In this study, we will compare WRF's simulations of tropical cyclones with those of the current generation of models (represented by MM5). Numerical experiments are conducted using MM5 and WRF using the same grid spacing, initialization, and comparable physics parameterizations for tropical cyclones of varying strengths in the West Pacific and Atlantic. The question we wish to address is, does WRF's improved numerics and physics lead to improved tropical cyclone predictions or is initialization still the dominant problem in numerical forecasts of tropical cyclones?