[plinkplonk]

Textbooks are about the only way to learn deep comp sci for someone (like me) who is self taught .

I don't know about the Levitin book, but the Cormen book is likely to be a little tough for someone just starting out. I would suggest Tardos and Kleinberg or Skiena's book (and once you are finished with those you can always come back to Cormen). Make sure you are comfortable deriving proofs before attempting Cormen. A lot of the value of the book is in its exercises and a lot of the exercises involve proving things.

So yes textbooks help, but in my experience they need to be the right textbooks in the right sequence. Good Luck!