Previous studies have demonstrated the capability of the European Centre for Medium-Range Weather Forecasts (ECMWF) 51-member, 32-day ensemble to forecast tropical cyclone (TC) events (formation and tracks) in the western North Pacific on the extended range (5–30 days). In this study, the performance of the ECMWF ensemble in extended-range forecasting of Atlantic TCs during May–December 2012 is evaluated using similar approaches. The conclusion from this evaluation is that Atlantic TC events have lower forecastability using the ECMWF ensemble than in the western North Pacific. Hurricanes Kirk and Leslie and Tropical Storms (TSs) Joyce and Oscar were successfully forecast in weeks 1–4 and, thus, are labeled as highly forecastable. Somewhat forecastable storms that are only forecast in three of the four weeks include Hurricanes Ernesto, Isaac, Nadine, and Sandy plus TS Florence. The limited forecastable storms that were successful in only the first two weeks include Hurricanes Gordon and Rafael plus TS Debby. The surprising result was that two hurricanes (Chris and Michael) and three TSs (Helene, Patty, and Tony) were not even forecast in week 1 before the starting time in the National Hurricane Center working best track (WBT) for these storms. As was the case in the western North Pacific, a substantial number of false alarm storms (no matches with any WBT) are predicted, with about 35% occurring in the first week. Except for the African wave–type false alarms, three other false alarm types may be easily recognized. A larger sample will be required to statistically verify the reliability of the probabilistic forecasts for the African wave–type ensemble storms.