In this paper, the revised RAPT algorithm and display are described and evaluated. The fidelity of the RAPT operational model is assessed by comparing RAPT departure status with observed departure flows (i.e., trajectories, weather avoidance maneuvers and storm penetrations) on several days when convective weather SWAPs were in effect in New York. Real-time in-situ observations at RAPT facilities (described in a companion paper at this conference; Robinson, 2008), user feedback from RAPT playbacks and the REPEAT web site are used to support this post-event evaluation. For example, real time observations provide the time and operational rationale for a specific departure route closure identified in the traffic flow analysis. This information is necessary to identify closures or flow restrictions that are the result of factors outside of the current RAPT algorithm domain (e.g., traffic restrictions due to volume, downstream congestion, etc.). Real time observations are also used to identify specific times when critical, weather-related operational decisions were made. The RAPT guidance at these critical decision points is analyzed to determine if RAPT provided information that enabled (or could have enabled, had it been used) more timely or effective decisions. The effect of forecast uncertainty on RAPT performance is also examined, particularly in convective weather situations where the location, severity and operational impact were difficult to predict. Strategies that mitigated risks associated with forecast uncertainty are presented. These include the use of additional information provided in the RAPT display, such as echo top heights encountered along the departure route, to confirm or modify RAPT guidance and the consideration of the departure status of two or more adjacent routes to 'average out' variations in the departure status timelines.