Good question. There is a spring predictability barrier (that we've written about before > https://www.climate.gov/news-features/blogs/enso/spring-predictability-… ) that makes it difficult for forecasters to know what will happen with a ton of confidence next fall/winter. 

This is one reason why using probabilities is so important. The uncertainty caused by this barrier is taken into account by forecasters when they are making their enso outlook