Using Release 2 of COADS we have constructed the ocean surface pressure, wind, wind stress and temperature composite anomalies for the Warm and Cold Events since World War II, and evaluated both the statistical significance of the month-to-month composite anomaly features and the extent to which each large-scale composite anomaly feature is typical in each event (Harrison and Larkin, 1996, 1998; Larkin and Harrison, 1999). Some of the issues surrounding the identification of Cold Event periods remain unsettled. Although there are many similar aspects in the equatorial Pacific, we find very interesting differences between the Warm and Cold composites away from the equator. In particular, the now-familiar "S-I" regression mode based on use of a cold tongue index is found to mislead in some significant respects about the 'typical' anomaly features to expect during Warm or Cold events. We contrast the features of Warm and Cold Event periods using their composites, and compare these with the regression approach to summarizing recent S-I variability in the Pacific.