Is There a One Best Method for Neural Networks?

For most of the history, the classic methods were really a good default way to go. If you have lots of data, this model and bit seems to be also a very good defalt choice. And if you are in low data regime, with high seasonality and some sort of easy trance, then i would say, the expolential, smooth things all the way to go.