This paper investigates whether the 2008 crisis in Korea was predictable by applying the signal approach. It shows that the early warning model constructed to predict the 1997 crisis in Korea works very well to predict the 2008 crisis implying the similarity of the two crises. In predicting the 2008 crisis, crisis period is defined two ways. One is to define the whole year of 2008 as the crisis period. The other confines the crisis period to the post September, 2008. The early warning was not sensitive to the crisis period. Both models provide early warning successfully. Well before the crisis period, signals increased rapidly and the conditional probabilities of crisis rose sharply. The most important indicators were the foreign debt/foreign reserves ratio and the foreign debt/total debt ratio of depository corporations. The models performed well in the out-of-sample period. It seems that Korea did not utilize the early warning systems to be ready for the 2008 crisis, although she had built them after the 1997 crisis to be ready for the another crisis. We must overcome the this-time-is-different syndrome and trust the early warning models to rely on early warning.