I've had a significant operation recently and been recovering, so haven't had much time to concentrate on chess (though I did get the coverage of the King's Gambit completed earlier).
I have a couple of deeply annotated games in the Sicilian Najdorf with 1.e4 c5 2.d4 cxd4 3.Nf3 d6 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 that I played recently.
Note the fairly unusual move-order. This move-order probably won't be right for most players: it works well if you're happy to play the White side of the Morra Gambit or the ...d6 lines of the Open Sicilian (while side-stepping lines like the Kan, Taimanov and Lowenthal). The main issue with it is that after 3...a6, 4.c3 is probably best, leading to a Morra, since after 4.Nxd4, after a subsequent ...e7-e5, White doesn't have the b5-square available for the knight on d4, and that after 3...d6 (as played in these games), 4.c3 Nf6 is awkward since 5.e5 (probably best) is now met by 5...dxe5, rather than a transposition into normal c3-Sicilian lines.
These two games highlight the downside of getting involved in these sort of highly theoretical tactical lines- at club level most of us don't really know what we're doing! But they do tend to produce pretty interesting games.
Of course, 6.Bg5 is one of the most "theoretical" responses to the Najdorf and there are plenty of ways of playing the Open Sicilian with White that are not as theoretical, though constructing a full Open Sicilian repertoire that both avoids heavy theory and maintains good attacking chances is quite tricky- some compromises will be needed one way or another against certain lines. For those who are interested, Michael Goeller at his Kenilworthian blog, back in January 2010, suggested a relatively aggressive and low-theory approach, mostly involving early f2-f4 advances, but also see the comments section at the end of the article for an illustration of the challenges involved.