Rank position forecasting in car racing is a challenging problem, which is featured with highly complex global dependency among the cars, with uncertainty resulted from existing exogenous factors, and as a sparse data problem. Existing methods, including statistical models, machine learning regression models, and several state-of-the-art deep forecasting models all perform not well on this problem. By elaborative analysis of pit stops events, we find it is critical to decompose the cause effects and model them, the rank position and pit stop events, separately. In the choice of sub-model from different deep models, we find the model with weak assumptions on the global dependency structure performs the best. Based on these observations, we propose RankNet, a combination of encoderdecoder network and separate MLP network that capable of delivering probabilistic forecasting to model the pit stop events and rank position in car racing. Further with the help of feature optimizations, RankNet demonstrates a significant performance improvement over the baselines, e.g., MAE improves more than 10% consistently, and is also more stable when adapting to unseen new data. Details of model optimization, performance profiling are presented. It is promising to provide useful forecasting tools for the car racing analysis and shine a light on solutions to similar challenging issues in general forecasting problems.