End of the day, you're never going to find any good books on this subject whatsoever and no one's going to teach it to you. I think some of those books could be approached as building a toolset in order to research strategies. However in actual design there's reasonable you'll ever find. If you want to learn how to design strategies, I suggest backtesting as many ideas as possible and possibly even executing some of them. If you lose money, I assure you'll quickly learn where you went wrong. There's no better teacher than a proper beating in financial markets..
Alright then what statistics theories do you believe in to let you believe your strategies would work in real trades? Welcome to suggest a good book for this as well.
Gloria, Are you asking which theories could be applied to developing new strategies? Or which theories are used to backtest existing strategies?
If you already understand basic data analysis (e.g. You can answer what the difference between data mining, overfitting, and data snooping is), then the rest is an art which isn't really taught in books. There isn't really a catch all, all strategies will have different requirements and beyond the math you'll run into considerable practical implications which are never described in books either. So without knowing more, it's very difficult to say what you should be looking at.