I think I have found how I might play the KID Samisch as Black. 6...c5! is a pawn sacrifice that has put some White players off playing the Samisch. The diagonal for the fianchettoed bishop is good, Black has a lead in development, and White's pawn on f3 will further hinder his development. White has a hole on d4, and furthermore, his king can come under danger. This would all be terrific compensation, but White also can eventually play Nd5. Still, I like this for Black.

In the Samisch attack, there is a certain gambit which is seen a great deal, that arises in this position:

click for larger view

White now has the opportunity to play 7. dxc5 dxc5 8. Qxd8 Rxd8 9. Bxc5 and be up a pawn, but with considerable development difficulties. Play often continues 9...Nc6 10. Nd5 Nd7 reaching this complicated position:

click for larger view

In an old copy of MCO, Nick DeFirmian recommends 11.Ba3! (his exclam, not mine). He was probably basing it on Beliavsky vs Chekhov, 1980, but let me tell you that was just about the last time White ever won that variation.

I found myself in exactly this spot against a young Chinese kid, maybe 12 years old, who totally ripped me apart. (To save face I must add that I saw his face in Chess Life with a giant trophy, I think this kid is 2000+ now.) Ever since that game I simply vowed to never accept that "free" pawn and refuse to play into their prepared lines.

click for larger view

Clearly 8...Nh5? 9.g4 is right out. If Black tries to simplify with 8...Qxd1+ 9.Rxd1 Nfd7 10.f4 then White is pretty much up a pawn for nothing. Clearly .

And so, we're looking at the natural response 8...Nfd7 and now 9.f4 of course,

click for larger view

Statistically this line does great for White (see http://www.chessgames.com/perl/ches...) but I don't pick my moves based on statistics. The fact is that White doesn't have nearly the same problems as in the other variation that I described.

Now I see that 9...f6 is common and logical as well, but White's game practically plays itself: 10. exf6 exf6 11. Nf3 Re8 12.Qd2 Nc6 13.O-O-O and I can't imagine anybody who wouldn't be happy with White's position here:

click for larger view

A great point of this variation, which makes me want to include in my repertoire, is that the moves are very easy and obvious for both sides. In the above variation, I'm not going to have to strain my head to remember to play 11.Nf3, 12.Qd2, or 13.O-O-O am I?

But wait, one more variation must be considered. The natural 9.Nc6, my computer's pick in the position as well. This might be Black's best option. But after 10.Nf3 I think we can stop looking at variations and make our "static evaluation" of the position:

click for larger view

I'll award this a easily. Why is White better? Three reasons IMO, #1 more space (even a beginner can see that it looks like White is pressing Black backwards) #2 better development (the same number of pieces, but White's are more active and useful.) #3 the pawn structure favors White. White can dominate the d-file with queen+rook(s) while Black really has no good files to place his rooks.

Looking at the games you give I wonder if Black should take on f6 with a piece.

In some lines e6 looks tempting for White but Bronstein vs Smirin, 1994 shows that this has some disadvantages.

click for larger view

The line i'm interested in goes 6.Bg5 c5 7.d5 e6 8.Qe2 exd5 9.Nxd5

click for larger view

which i think gives white a good position. Play often continues 9...Be6 10.Ne2 Bxd5 11.cxd5

click for larger view

white then puts the king knight on c3, develops the kings bishop to e2 and castles. I'm wondering if avoiding castling and going for h4 at some point is a good idea? Blacks king looks very safe and it seems one way to break open that position. White also has more space and the bishop pair, although he needs to complete his development.

Looks to me like Black has a good Benoni after 11...Re8, preventing the exchange of dark bishops after Bh6.