These latest comments raise another interesting point. Since I've never used a database nor an engine to select moves, I mistakenly lumped them together. I stand corrected, as apparently, an engine is not allowed because it finds moves for you. Whereas a database merely shows moves that grandmasters played in similar positions, and it's up to the player to do his own research. Do I have the distinction correct, or am I still missing the point?
Whatever the answer, I still like my strategy of play first, analyze after. But I'm only an annoying idealistic beginner.
.
htdavidht: I think you overestimate the impact that using an opening reference has. It generally doesn't give you an advantage over your opponent (especially if they are also using one) - it just (hopefully) helps to ensure that you reach a balanced and playable middlegame. You still have to be able to play that middlegame and any resulting endgame in order to win.
Also, if your opponent goes out of book, you usually have to work out the refutation yourself - the database won't always tell you that.
If my oponent goes "out the book", most likely that would mean he blunders, unless my oponent is some Master and is creating a new line on the body of opening theory.
I have been faced with a reasonable novelty as early as move six.
But in most of my games novelties appear around moves 9-15. And they are not blunders (And I´m not speaking about the chess.com database that it´s almost empty).
But sure, I don´t play complex systems like the Najdorf or the Botvinnik Variation in the Semi-Slav.