Since the mid-1970s, ENSO has changed in character to a predominance of El Niño conditions, the extreme phase of which appears coincidental with increases in global temperature records. Instrumental time series (<150 years) are too short to adequately address the significance of late twentieth-century ENSO variability, thus, multi-century palaeoclimate reconstructions derived from long proxy records are sought. Despite the global influence exerted by ENSO on society, limited consensus exists within the scientific community as to which index best defines the timing, duration and strength of events. Here we address issues associated with the complexity of ENSO characterisation by comparing the ‘event capture’ ability of two currently used indices of ENSO. It is suggested that the use of a sole ENSO index is undesirable as a given index is only indicative of one physical aspect of the phenomenon, and as such is unlikely to be representative of the wider interactions experienced in the coupled ocean-atmospheric system. In an attempt to describe more of the nature and evolution of ENSO events, the Coupled ENSO Index (CEI) classification scheme was devised to identify synchronous oceanic (Niño 3.4 SST) and atmospheric (Southern Oscillation Index) anomalies associated with ENSO for the instrumental period (1871–2003). The CEI is of practical relevance to the ENSO community as it provides an amplitude preserving instrumental baseline for the calibration of proxy records to reconstruct both components of the ENSO system. Analysis of the nature of instrumental ENSO events from the CEI suggests that the frequency and intensity of post-1970 ENSO events (when 50% of all extreme events identified occur) appears the most anomalous in the context of at least the past century. It is hoped that the CEI will facilitate palaeo-ENSO research to systematically resolve the long-term context of past ENSO behaviour to assess whether the apparently anomalous nature of late twentieth-century variability is unprecedented within existing palaeoclimate archives. Copyright © 2005 Royal Meteorological Society.