King's Indian Mar del Plata 9. Nd2 busted?!

Although the Mar del Plata variation of the King's Indian is no longer as popular as it once was, it is still played rather frequently even at the top levels of chess. In the main line of this variation, Nd2 is the third most popular white move and has been played by strong contemporary grandmasters such as Viswanathan Anand and Boris Gelfand. Nevertheless, another chess.com user, birdsopening, and I believe that this line is actually busted for white because of the responses, 9... Nd7! and 9...Ne8! If black utilizes his attacking resources properly, white should be deprived of all possible winning chances and black should win and at worst draw every time.

At first, the notion that two USCF ~2000 patzers could possibly bust a line that has been used by players of the caliber of Tigran Vartanovich Petrosian and Viktor Korchnoi seems utterly ludicrous. However, most of the legwork was actually done by much stronger players than ourselves such as Hikaru Nakamura, Pascal Charbonneau and Irina Krush, to name a few. We are simply compiling their efforts and contributing some of our own ideas.

Both 9...Nd7 and 9...Ne8 look incorrect. They allow white to execute his queenside plan unhindered unlike the more popular 9...a5 or 9...c5 and don't score as well as the alternatives. In theory, white's attack should be faster than black's, but this is not the case provided black plays accurately and utilizes his defensive and attacking resources properly.

The first line that birdsopening and I analyzed is the main line, where white chooses his most sensible queenside plan, which progresses as follows:

While dxc5 seems counterintuitive, willingly undermining the pawn chain and allowing white to create a passed pawn with an eventual b6 break, it is the best move in the position. Anything else scores well for white and for good reason - anything else allows b6. dxc5 slows down white on the queenside enough to create a dangerous kingside initiative:

The diagrams aren't working... how do I fix them?

EDIT: Never mind, I inserted them again and they work now.

didnt read the text, if i did i probably would have stopped at we have busted a line...

but that being said does anyone know a book that teaches french the korchnoi way??

 anyway analysing is good, keep it up.

It is, although, if I recall correctly, the Mar del Plata variation denotes any variation of the King's Indian with 5. Nf3. The mainline of the Mar del Plata with 8...Ne7 is also sometimes simply referred to as the Mar del Plata variation, but the line is technically called the Aronin-Taimanov variation. On a side note, I'm glad to see that you that Kf7 is an effective attacking idea, and thanks for your input. 

@Tacticator: I mean no disrespect, but your comments make it evident that you have little knowledge of Mar del Plata positions. Of course you're supposed to play on the side of the board where you are stronger, but g4 is not a bad move and is a widely accepted defensive idea that has been employed by various strong players, Kasparov included. And, contrary to your assertion, it shouldn't always be met with fxg3. H5, opening the h-file, is also good. Furthermore, dxc5 is very good for black. 18. d6 is met by Bf8 and black wins the d6 pawn. Dogmatism is acceptable as long as you know what you are talking about; in this case you don't.

I apologize If I sound rude, but I usually do on the internet.

Amazing work RRM888.

I wanna say thank you even if I don't play these lines :)

I don't know theory on the KID but looking at your second diagram at move 19 it seems to me that white is just being too careless with his kingside and missing simple defensive ideas. That rook on f1 is hindering white's coordination, it should have been moved. On f1 there should be a bishop (to allow to answer g3 with h3), or a knight that would offer a better protection of h2.

I don't know theory in these lines so please correct me if I'm talking trash. I'm just saying what I istinctively would have done.

I am a proud King's Indian Defence player of USCF 2000, and I approve RRM888's message. 

A rare moment of lucidity. What does one (1) game won by Black proves, exactly ?

Back to chess : you dismissed Tacticator's idea of 14...fxg3 (ep) in the second game, I would like to know why, apart from "someone played it and obtained a bad position". For instance 14...fxg3 15.hxg3 g4 and the kingside blows up.

Uhm...NO!

The Mar Del Plata is 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 O-O 6.Be2 e5 7.O-O Nc6 8.d5 Ne7.

5.Nf3 doesn't constitute the Mar Del Plata.  For instance, 6.h3 is a completely different line that also starts with an M but I can't recall it.  7.d5 instead of 7.O-O is the Petrosian Variation while 7.Be3 is the Gligoric Variation.  7...dxe4 is a sideline for Black while 7.dxe5 is dubbed "The Exchange Variation".

Thank you, ThrillerFan, for clearing that up and I believe that the h3 line is known as the Makagonov variation.

@IronTiger: Of course one game proves nothing, but our analysis is comprised of more than just one game. We compiled a number of ideas that have been used before in grandmaster games and added some of our own ideas. If you believe our analysis is incorrect, please tell us where we went wrong.

I worded my response to Tacticator badly. I didn't mean to completely dismiss fxg3, but instead meant to say that it was not, as he asserted, the best move in the position. I believe h5 is better as opening and lining up pieces on the h-file appears to be far more dangerous than fxg3 followed by g4, which allows resources such as Be3 and Bf3 which help defend the kingside. We haven't done much concrete analysis on the line so I can't say for certain, but h5 appears to be better for black, although fxg3 looks playable. 

Of course White has resources after ...fxg3 and ...g4, but ...h5, though it opens the h file, blocks the f and g file which means a considerable limitation of the attacking potential. I'm not saying it does not open the file : I'm saying it blocks them, which is different, and in a way that does not make sacrifices promising (like the thematic ...Bxh3 after ...g3 h3 in some lines). After ...h5, as far as I can see, White can just oppose a rook on h1, put the king on g2, clear the back rank and both rooks do a great job at fighting the h file. Black cannot immediately put his rooks on the h file either, and you can delay ...hxg4 but remember White just has to put king on g2 and rook on h1 which is not a huge amount of material - the rest is operating on the queenside - when you need at least two rooks, so White will win the hide-and-seek game.

On the other hand, after ....fxg3 and ...g4, the kingside completely blows up, which will give Black some play on that side. Maybe not enough play ; maybe White can, actually, defend adequately and/or even take the initiative on this wing ; but in any case, it looks better than your suggestion.

The famous game from recent years in the 9... Ne8 line is of course the Gelfand-Nakamura from the world team championship 2010 given in RRM888:s second diagram. There was a later game Kramnik-Nakamura in chess olympiad 2010 and although the game ended in a draw the concencus among the commentators seems to be that Kramink had the superior game. Nakamura probably agrees as in a later game against Kramnik in Amber rapid in 2011 he switched back to the modern mainline 9... a5. By the way, Kranik doesn't normally play 9. Nd2 and one can speculate that he only chose it because he felt that Nakamura's handling of that line isn't quite correct - a strategy that obviously backfired as he only scored half point out of the two games!

After fxg3 and g4, the kingside hardly "blows up". White's kingside pawns are disturbed, but that does not translate into an attack and black has zero compensation for the pawn as well as a very unpleasant position:

H5, on the other hand, gives black an actionable kingside plan and gives him good kingside play. After piling up on the h-file, black can truly make the kingside explode after exchanging on g4 and perhaps sacrificing a piece on g4 or a pawn on f3(or both). I reiterate that there is much analysis to be done but it looks far more promising than fxg3. 

Yes, but Black can wait before playing ...g4. For instance 15...h5.

Whereas after 14.g4, taking en passant is available for only one move.

15...h5 loses to Bxg5. Did you look at the diagram?

I looked, but I did not see. 

Indeed. Well, putting apart the ridiculous 15...h6, it means the only not-immediately-refuted try in that line is 15...Qe8 (with the idea 16.Bxg5 Qg6), but I do not trust it. If this fails too, I agree that ...fxg3 is wrong.

Otherwise, in the sideline 15...Nh5, I think 17...Nf4+ is a better try that ...Nh4+. But White should survive.

15... Qe8 drops d6. I thought Nf4+ looked promising but white is doing well after gxf4 regardless of how black recaptures.