Recently i became very interested in sicilian Najdorf and thus started learning it, this is one of the games i played with my friend, who also likes to practice sicilian. I promised to him make annotations and show that to him, but then i thought maybe other players would be interested too, so i am posting it here insteed of blog section. If you have any comments and suggestions you are welcome. Thanks in advance.

while you are very interested in the Najdorf, I have became obsessed with the Dragon. Im curious about the bad reputation it receives ('I have got it all down to science, sac, sac, mate...' - Bobby Fischer)...

Dragon is fine and dont refuted! Even that famous game Fischer-Larsen, that is first or second in Fischer's book "My 60 memorable games" - even there Larsen missed good move that totally brokes white attack and gives at least an equilizing and maybe some chances for attack for black. Yugoslavian attack has no forced wins, black still has a lot of lines and dirty tricks, and also you may switch to accelerated dragon in some cases, to avoid Yugoslavian. So if you are obsessed - play it, enjoy it and dont even doubt is it sound or not. Even Kasparov played it several times against best rivals and didnt lost it.

I believe that analysts claim after 1.e4 c5 2.Nf3 d6 3.d4 cd 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 Bg7 7.f3 0-0 8.Qd2 Nc6 9.Bc4, Larsen should have played Bd7 instead of Nxd4.

I actually have a book on the Accelareted Dragon, but some line which they recommend are just completely dead draws. Surely, the authors should understand that some players aren't always satisfied with a draw : )?

Very nice game!  And a very good example of how unforgiving the Najdorf can be if you don't have concrete knowledge of the various lines.  Often times, moves that seem perfectly natural come back to bite you as seen here with 11.a3 (for the reasons you pointed out) and 13.g5. 

The latter move is especially troublesome because at first glance it looks so logical - you'd never guess that it scores less than 40% compared to more than 55% for 13.f5.  The advantage of 13.f5 is that it encourages black to castle immediately which is something that white very much needs in order to make his kingside pawn surge count for something.  Example: 13. f5 O-O 14. g5 Ne8 15. Rg1 b4 16. axb4 Rxb4 17. Bh3 Qb6 18. f6 Bd8 19. Qd3 gxf6 20. gxf6+ Kh8. 

As you clearly illustrated in this game, if black has time to initiate queenside operations before castling, it becomes very difficult for white to prove that all of those pawn moves were justified.  Those are the sort of things that are very difficult to reason over-the-board....

Most books on theory are written with the supposition that both players will find the "best move" every time.  For every position, "best move" should end in a draw, since that is the natural tendency of any game (from game theory).

It's rare, even at the highest levels of play, for anyone to play perfect Chess, so a person who isn't "always satisfied with a draw" needs only for their opponent to make the wrong move... 

Catalyst: You play really good Chess!  I've always been crappy with the Sicilian, but I'm going to try the Najdorf now...

Then pick a Najdorf - that is deathmatch, no draws.

Right, and later when Fischer already attacked and already sacrificed the exchange - Larsen again had good equalizing move, that guarantees at least a draw. Do you know it?

...but also so interesting and adventurous!  Thanks for your input! If 13.f5 i wouldnt dare to castle right under direct attack, i would play 13...e5 and remain uncastled to make white spend more time for exposing my king, and for this time i would try to prepare something hot by myself as well. Seems like castling in Bg5 Najdorf is rare thing with black.

Something along the line of ...Rc4 by black at one stage... giving the exchange bak?

Exactly. Now your dragon is unbeatable, feed him well.

Heheh, you smashed me in that game Cata.  I did not think enough. :(

After my move 30. e5..I was hoping for you to play 30..dxe5. 31. Rd1 31..Qxf6 32. Nc7+ 32..Ke7 33. N7d5+ 33..exd5 34. Nxd5+ winning the queen for 2 knights.

But it was not to be.

white has advantage when Knight gets to D4 early to control board. It seems like black was sitting around for white to make mistakes to take advantage. Many great chess player and GM rarley make mistakes, so hoping for a mistake is just a mistake

But Im not a GM. ;)

Really? So your saying that the main-line is inaccurate and black could have played better? As he states, until 11. a3 everything played was book so I really don't see how you can insult his play. 

in this opening of sicilian for black, black has a poorly developed dark bishop, also black lets white gets its knight to d4 which controls whole middle and is opened nicely. White should of won if he made better use of his pieces to open up blacks defense. If you want to play the sicilian i suggest playing it in a different way. You guys should stop talking about book openings and computers because they will lose to humans who play very well. Think of your own openings and stragetgy and learn from your mistakes. For example certain pieces may be better to have in certain cituations even if you have to give up a valuable piece like queen or rook even if computer and books disagree.

So how should white have played? Do tell because it seems that you know a refutation to the sicilian and I would be intrigued to learn it. 

 you can see a game I posted to learn my version of the sicilian. It gives you a different look from book and computer openings.

hmm... computers having a different view of your material sac... i understand, as not all programs have adequate understanding of compensation. But books... I am yet to see one. Any decent book should have covered all the main line, and any of those sacs, because those books are written by humans, who actually does understand the meaning of compensation.

 incorrect, if you play the poisoned pawn variation as black(which is one of the critical variations) then theory says white can force a draw. 

Everyone check out this puzzle I created in which I finally beat Fritz11 at its highest rating.     http://www.chess.com/forum/view/more-puzzles/gm-middle-game-puzzle