See, if someone reads all of this analysis and understands it, the black player should be in some trouble! Vice versa, but white's getting the reward with a nice advantage almost winning. The move ...Nfd7 is more solid but less challenging to white's play. It's just harder for black to carry out his plan that way. White should play a4 with advantage because black's knight's will end up on d7 and c7(after...Na6), somewhat passive.

14. ... a6 is a mistake.

Better to try is 14. ... Qg2 which leads to an unbalanced but equal position. The trade of queens here is practically forced with 15. Qf3 Qxf3 16. Nxf3 fxe6 17. dxe6 a6 18. Bxd7+ Bxd7 19. exd7 Kxd7 20. Bxc5 Kc6 and the position is equal. There are plenty of open files for the black rooks to operate on and black posseses a kingside majority. 

It seems to me that most of that line is forced. Don't take my word for it because I sat here and played it out myself (I didn't look up my refutation in the MCO :) ). Let's discuss this position.

 That position looks relatively balanced, but I think after ...Qg2 instead of trading queens white could play exd7+ Bxd7 Bxd7+ Kxd7 and now I think white can just play Qa4+ Kc8 0-0-0 where c5 is weak, the passed d pawn can be very strong, and white has some chances for attack. If ...Qxg3 then Bf4 is very strong. The possible d pawn push looks brutal and with queens still on black has to worry about his king. White's king isn't ultra safe, but he's much more mobilised so I think he has the initiative.

No, black is better in that position. 

14. ... Qg2 15. exd7+ Bxd7 16. Bxd7+ Kxd7 17. Qa4+ Kc8 18. 0-0-0 Bxc3! 19. bxc3 Qxg3 and black is better here with the prospect of having three connected passed pawns. Black will aim to trade queens and then push on the king side. I like Blacks position here.

So far I haven't seen any refutations of the Benoni...

I just put in all the moves in fritz to see what it thought and it said white was up 1.5 after Bxc5. Black has a huge kingside majority, but white's central pawns are more threatening since they immediately require black's attention and white's extra minor pieces are actively in play even if black can trade queens. Black is considerably behind in development so I'm not suprised if white can force serious threats. So white should avoid that endgame position you posted, though there too I think white is better but just slightly because his pieces at the moment are superior to black's so he has time and the only thing black has in compensation is his kingside majority.

Yea...you're right. It took Fritz a while on my computer before it recognized White to have any advantage at all. I suppose the line black may want to try is:

20. Bxc5 Qxc3+ 21. Kb1 Qxc5!? 22. Rc1 Qxc1+ 23. Kxc1 which is slightly better but White is still better....

I don't know then...I guess this is why I play e6 first haha...I never really thought that White had this much of an advantage in this variation. I stand corrected.

In the CC universe, black has scored some impressive wins with 8...Nfd7.  Below is an upset win from a recent WC cycle.  I also included a couple other games that looked interesting.  It's pretty clear from scanning through high level CC games that 8...Nfd7 has been the most successful move (by a substantial margin) and is the preferred choice of Benoni specialists like Mueller who's racked up a lot of points with it the last few years.

I have chalked up a couple of wins as black in the 8...Nfd7 line, which I take to be the main line these days.  I'd be interested to know what the refutation of that is.

No, after 20 Bxc5! White is clearly better.  Black may have 3 connected passers, but White's advanced passer on d5 has more immediate influence on the game.  Black's King is poorly situated and quite vulnerable - if 20 ... Rd8  21 Bb6 is possible, or just 21 Qd4 threatening a7 amongst other things.  Of course 20 ... Qxc3+ looks awkward after simply 21 Kb1, with Rc1 to follow.  Trying to push the pawns by 20 ... h5 gets ugly fast after 21 Qd4, and the attempt to defend with 20 ... Qc7 also meets 21 Qd4 Rd8 22 Nf3, and once White's N enters the game, his threats will multiply.

The earlier variation with 15 Qf3 is weak.  White definitely wants to keep the Qs ON the board.  When you have the general disposition of material of two minors against a Rook, the advantage is in attacking, and for that the Q is most helpful.

The ...Nfd7 line hasn't been refuted like ... Nbd7, but it also leads to difficult positions to play for Black against a generally persistant White initiative.  It isn't unsound, and may be better to play against a White player who is used to playing against the KID or Grunfeld and less familiar with the niceties of the Benoni structure than entering the same KID main line contested last week between Ivanchuk and Gelfand, which he has studied.

It's not unplayable.  But against a player familiar with it, it could be uncomfortable.

Well 8...Nfd7 is not really the move black wants to play and after moves like 9 a4 Black can only get in ...b5 via ...Na6-c7 where white's development is more natural and may be able to get in e5 well by that time. This is just my opinion but I have this on chessmaster and Seirawan prefers white's game moderately. Guess what, after the end of the unambitious (but forced) ...Nfd7 is when I abandoned the benoni! And what is the move order that aviods the taimanov? My idea would not be an early ...e6, but I was thinking a king's indian move order where white is likely to play Nf3 and then I play ...c5 expecting d5 then ...e6. I think this is the best way for black to play the benoni and although white is slightly better I admit it is a viable variation for black. Maybe white could still avoid the benoni though by ignoring the c pawn and play for a sort of sicilian maroczy bind instead though.