It sounds like there is a difference of opinion on when "early" is. It is well known that forecasts made in the spring for ENSO have less skill then those afterwards. This is known as the Spring Predictability Barrier which we have written about here . So we know that forecasts made during that period are not as good. (Plus as a note, the forecast you showed above in your link is a model-only forecast and not the human based forecasts that are issued at the beginning of every month). So I'd recommend looking at forecasts after that period, which would still be 6 months before the upcoming winter.

Of course, the other issue when determining how good a forecast is, is that our forecasts are probabilistic in nature. These sorts of forecasts need many events in order to see just how good they are. After all, if we issue a forecast of a 60% chance of El Nino and a 25% of La Nina by DJF and it ends up being a La Nina, is that wrong? That is supposed to happen 25% of the time.