Over the last six or seven years, seasonal forecasts provided by the Bureau of Meteorology's National Climate Centre have been based on statistical relationships between rainfall and the EI Nino-Southern Oscillation (ENSO) phenomenon. The Southern Oscillation Index (Sol), a measure of the atmospheric pressure differences between Darwin and Tahiti, has been used extensively as an indicator of ENSO events. Significant correlations with the Sol, however, are limited to certain seasons and locations only, and strong relationships are often only found once an ENSO event is well developed. In particular, correlations over the period from late spring to early autumn in the cotton growing areas of eastern Australia are generally quite poor. Forecasts have also been restricted to 3 month seasonal average values commencing immediately from the time of issue. Some improvements in statistical techniques using trend and phase information has extended the use of Sol based forecasts to provide coverage for most of the year round, and useful assessments of overall conditions are provided, but for cotton growing areas the skill has not increased to the point of bulletproof reliability. What then can be expected in the future?