On 7 January 2014 an X1.2 flare and CME with a radial speed ≈ 2500 km s −1 was observed from near an active region close to disk center. This led many forecasters to estimate a rapid arrival at Earth (≈ 36 hours) and predict a strong geomagnetic storm. However, only a glancing CME arrival was observed at Earth with a transit time of ≈ 49 hours and a K P geomagnetic index of only 3−. We study the interplanetary propagation of this CME using the ensemble Wang-Sheeley-Arge (WSA)-ENLIL+Cone model, that allows a sampling of CME parameter uncertainties. We explore a series of simulations to isolate the effects of the background solar wind solution, CME shape, tilt, location, size, and speed, and the results are compared with observed in-situ arrivals at Venus, Earth, and Mars. Our results show that a tilted ellipsoid CME shape improves the initial real-time prediction to better reflect the observed in-situ signatures and the geomagnetic storm strength. CME parameters from the Graduated Cylindrical Shell model used as input to WSA-ENLIL+Cone, along with a tilted ellipsoid cloud shape, improve the arrival-time error by 14.5, 18.7, 23.4 hours for Venus, Earth, and Mars respectively. These results highlight that CME orientation and directionality with respect to observatories play an important role in understanding the propagation of this CME, and for forecasting other glancing CME arrivals. This study also demonstrates the importance of three-dimensional CME fitting made possible by multiple viewpoint imaging. 9 In this coordinate system the Z axis is aligned with the solar north rotation pole and the X axis pointing toward the intersection between the solar equator and the solar central meridian as seen from Earth (Hapgood 1992;Thompson 2006). HEEQ coordinates are related to Stonyhurst heliographic coordinates, with directions south of the origin represented by negative HEEQ latitudes, and directions east by negative HEEQ longitudes.