I don't know much about Soros or Loeb, but Livermore, Darvas and O'Neill was all fairly methodical as far as I can remember.
And that's the key. There needs to be a methodology and some kind of consistency in your approach. IMO, at least. And how to best build a methodology? Well, live trading...