Columns | August 15, 2009 17:50

Finding Nepo (on an old laptop)

chess.netFor over a decade, I have been ridiculed by chess friends for playing the Centre Game, an opening with a respectable history which nevertheless has a very bad image in modern opening theory. In Mainz, I witnessed Ian Nepomniachtchi and Levon Aronian play a very exciting game, reminding me of a time when internet chess had just begun and I had the golden opportunity to test the Centre Game against some leading grandmasters, including Alexei Shirov.

First, though, a little apology. It has come to my attention that Ian Nepomniachtchi isn't a big fan of the nickname 'Nepo', so I won't use it in this article. But I hope he can appreciate the little pun in the title of this article, which is not exactly about finding Nepomniachtchi himself, of course, but rather his ideas in the Centre Game.

I remember logging onto the server with telnet for the first time in 1996. I already had some experience with other online chess servers, but was the new kid on the block in those days, featuring a host of very strong grandmasters who also played against amateurs like me from time to time. I had met two of them, Alexander Shabalov and Alexei Shirov during the Donner Memorial tournaments in Holland.

As a matter of fact, it was Shabalov who had told me about the Centre Game for the first time, leading me to analyse this opening and make it part of my repertoire. On, I got the opportunity to play quite a few practice games against them - and in all games where I was white, I tried playing the Centre Game. Sometimes, this was not so easy because at home I still only had an ASCII-based interface, forcing me to execute the moves on a real board next to my monitor first and then typing them on the screen...

But enough nostalgia already. The reason why this story is interesting at all is because after watching Nepomniachtchi vs. Aronian, I was inspired to dig up my old laptop and look up some of the the games that I had saved there. While replaying some of them, I realized many of these games we played back then - and their subsequent analysis - are still relevant for current opening theory. In this article, I want to present some of these games to shed some light on the recent Mainz game. And we will find many traces of Nepomniachtchi's recent experiment in those blitz encouters - light-hearted as they sometimes are.

Looking through these 'test games' again, as well as the available games in databases, I noticed most of them imply that Black has to make three crucial choices in this variation. The first is the choice of retreating his bishop: Bb4-f8 or Bb4-a5. The second is where to go with the e4-rook. The third is whether to play d7-d5 or d7-d6. And all three decisions made by Aronian are very revealing.

Let's first see how the opening moves went and make ourselves familiar with the most important positions.

Ian Nepomniachtchi - Levon Aronian
Mainz (rapid) 2009

1.e4 e5 2.d4 exd4 3.Qxd4 Nc6 4.Qe3 Nf6 5.Nc3 Bb4 6.Bd2 0-0 7.0-0-0 Re8 8.Qg3!
dia1Most black players know that this move, which was first played by Frank Marshall, is White's most dangerous idea in the Centre Game. In my opinion it is a very logical idea: Black doesn't have many defensive resources on the kingside, so White immediately takes advantage of this by creating direct threats against Black's king. Of course, this comes at the cost of a pawn, as is well known:


This move is almost played automatically, at least in my experience. But 8...Nxe4 is also interesting. After 9.Nxe4 Rxe4, Nepomniachtchi himself played the old move 10.Bf4 against Vallejo Pons in Moscow, 2007, but after 10...Qf6 11.Nh3 d6 12.Bd3 Nd4! 13.Kb1? (13.Be3 occured in the ancient game Mieses, Capablanca 1913, and now the typical move 13...Rg4! is good for Black) 13...Ne2+ 14.Kb1 Ba3!! dia2
Black was virtually winning already. However, after 10.c3! things are much less clear. I believe White has dangerous chances here, even though, of course, he is still a pawn down.

9.a3 This is the tabiya of entire Centre Game, and it was also this position which occured in most blitz games that I played with this line.  White is trying to keep the game as concrete as possible, and it is not so easy for Black to decide how to react. Black must now choose between three major plans:

1. to retreat the rook to e8, trying to consolidate before White gets his attack going.

2. to provoke even further complications by 9...Ba5! - sacrificing the exchange (or even more in some cases) for a direct counter attack.

3. to try to keep harrassing the white queen by means of 9...Rg4!? - Aronian's choice.

Let's check the first option. After 9...Bd6 10.f4 Re8 11.Bd3 the position is unclear, meaning relatively untested in grandmaster practice, but full of chances for White. (11.Nf3 Bf8 12.f5, as played in the game Shytaj-Gustafsson, Reggio Emilia 2009, is also interesting.) White's plan is very simple: to exchange rooks on the e-file, clear the way for the moves f4-f5 and g2-g4 and start a slow but difficult-to-stop attack on the Black king.

It's funny to look in the database for games after 11...Bf8 12.Nf3 d5 because they're all from 1998 or later. However, back in 1997 I had already faced this double-edged position in a blitz game against Larry Christiansen (where I played the dubious 13.f5?! instead of 13.Rhe1, after which I was unpleasantly surprised by 13...d4! and complications were in Black's advantage.). Christiansen was quite ahead of his time in realizing that in this case, the active move d7-d5 is better than the more passive d7-d6. But, as we will see, this is not always the case.

Interestingly, in this line, too, there is discussion about where to go with the bishop, although here Black can choose between 11...Bf8 and the riskier 11...Bc5. It's noteworthy that Shabalov, who has played this line with White as well, preferred 11....Bc5 in a number of blitz games against me. For White, plans remain roughly the same, trying to attack on the kingside with a pawn storm in both cases.

Option 2 is to play 9...Ba5
dia3in order to keep the pin of the Nc3 alive. Also, the bishop may be better on b6 than it is on, say, f8. This move was played against me by Alexei Shirov on several times. I believe it is the strongest possibility for Black in this position. Let's first of all establish that White can't win an exchange by 10.b4 because after 10...Bb6 11.Nxe4? Nxe4 f2 is hanging. Also, the normal developing move 10.Bd3 does not work in view of 10...Rg4 11.Qe3 d5 and Black is almost winning already.

It's clear White has to prevent this. The most logical try, then, is 10.f3, which Shabalov tried once, but unfortunately it is refuted by 10...Nh5! and White's position is quite terrible all of a sudden. So somewhere around 1997, together with some club members, I started analyzing the slow-looking  move 10.h3!? and then was lucky enough to be able to test it in a few blitz games against Shirov, who was interested in the whole line since he tried it once as white himself (against Karpov, Dos Hermanas 1994). Well, even though 10.h3 couldn't save the variation, at least it resulted in some very entertaining games.

One crazy game, which was played on ICC in 1998, shows why d7-d5 is not always the right way for Black: 10...d5?! 11.Bd3 Re8 12.b4 Bb6 13.Bg5 a5!? and now I thought I had him by playing 14.Nxd5!
dia4 - note that with a pawn on d6, this wouldn't be half as strong. I now thought I was winning, but Shirov remained true to his principles and played the amazing 14...Nxd5!! 15.Bxd8 axb4 and although White objectively has the advantage, I couldn't withstand the 'fire on board' and helplessly lost the game.

In the next game, it turned out that in this case, playing d6 is much better: 10...d6! 11.Bd3 Re8 12.b4 Bb6 13.Bg5 Ne5 14.Nd5 Nxd3+ 15.cxd3 Bd4 16.Nf3 Be6 17.Nxd4 Bxd5 18.Nf5 Be6
dia5and now I should have gone for a draw which could be obtained after 19.Nxg7! Kxg7 20.Bh4+ Kh8 21.Qf3 = (Shirov), but instead I played 19.Nh6+? which backfired after the simple 19...Kf8! and Black was winning already. On top of that, 18...Re5! leads to a clear advantage for Black.

With 10.h3 being more or less refuted, it's unclear what White can still try. An interesting attempt is 10.h4 (trying to hide the queen on h2 after ...Rg4) but, again, 10...d5! seems difficult to counter for White. I don't really believe in the compensation after 11.f3 Re8 12.h5 Nd4! as in Paragua-Korneev, Seville 2002. Unless Nepomniachtchi had prepared something in this line, I think Aronian should have gone for 9...Ba5!

Anyway, Aronian played the move 9...Rg4 dia6in Mainz.  This move, too, was also tried various times in the test games I played in those early internet days. It has long had a reputation of being a good way to counter White's opening idea. But after seeing the Mainz game, I am not so sure about this anymore.

10.Qe3 Bf8! In the 90s, I still thought that this was inaccurate, but now I think this is a very important improvement over 10...Ba5. Moving the bishop to a5 looks just as logical here as it does on move 9, so why is Bf8 still better here? Well, there's a catch with the bishop on a5, as becomes clear when we look at another blitz game I played against Shirov.

After 10...Ba5 11.f3 Rg6 12.h4! Black runs the risk of losing the exchange, as happened in the rapid game Shabalov-Computer Socrates, Boston 1994: 12...Bb6 13.Qe1 d6 14.h5! Nxh5 15.Rxh5 Bxg1 16.Bd3 attacking both g1 and g6, and White has the advantage if he takes on g6 (which Shabalov didn't, by the way.) So, Shirov tried the sharp and straightforward 12...d5 13.h5 d4 against me in one blitz game, and it worked because I went bezerk with 14.hxg6? dxe3 15.Bxe3 Qe7 -+ Moll-Shirov, blitz 1997). However, the cool 14.Qe1! refutes Black's idea, because 14...Nxh5 fails to 15.Rxh5 dxc3 16.Bxc3.

dia7Finally, we see one of the disadvantages of having a bishop on a5: sometimes it hangs! In a later rapid game Kislinsky-Kukawski, Warsaw 2003, Black played 14...Rh6 instead of 14...Nxh5, but after 15.Bxh6 White won easily. I can't help wondering whether Aronian saw this idea, inspiring him to play Bf8 instead of Ba5. In any case, it was already known to insiders since 1997!

All this seems to indicate that Black cannot always give up the exchange so easily, and his best try may well be 11...Rd4 with some simplifying threats connected with Bxc3 and/or Nd5. Back in the days, the computers assessed the position after 12.Bd3 Bxc3 13.Bxc3 Nd5 14.Qd2 Nxc3 15.Qxc3 to be good for Black, but in an article I wrote for the German magazine Kaissiber in 1999, I pointed out that White in fact has good compensation after e.g. 15...Qg5+ 16.Kb1 b5 17.h4! and Black's position is surprisingly unstable.  Also, after 15...Rh4 nowadays' engines see very decent compensation for White after 16.Ne2 d6 (16...d5 17.Be4!) 17.Be4!? and White has active play for the pawn, quite in spirit with what he initially hoped for when embarking on 8.Qg3. (In the game Campora-De La Paz, Decameron 2003, White won after 17.h3!? with the idea to shut the rook off with g2-g4.)

We can conclude that Black definitely has problems after 10...Ba5, so Aronian's move is a very important one!


Alexei Shirov - who discovered the novelty 11.f3! already back in 1997

11.f3!N This logical move is considered to be a novelty - and a very good one it is too - but it was already indicated by Shirov in 1997 as an interesting alternative to h3, which somehow looks less logical. Admittedly, after 11.h3, too, the position remains highly unclear due to 11...Rg6 12.Bd3! Rxg2 14.Nge2 as happened in Xie Jun-Flear, Hastings 1996. (After 11.h3 the move 11..Rd4 is rather risky in view of the simple developing move 12.Nf3). After the text, Black now has a third big choice to make: where to play the rook?

11...Rg6! Again, a brave choice by Aronian! As is clear from the above comments, the black rook is often lost on g6, and this is also the case in Aronian's game, but he gets very good compensation for it in return - even though the computer doesn't agree with this. The 'engine-inspired' alternative was 11...Rd4 with the idea Rd4-d6-e6, solidifying the position, but this is very risky, especially if you're human instead of a silicon monster. True, after the sloppy 12.Bd3 Rd6! I don't think White has enough compensation for the pawn, but 12.Nge2 poses the rampant rook for new problems: where to go now again? If 12...Rd6, then 13.Nf4 (preventing Re6) 13...Nd4 14.Bd3 Ne6 15.g4 still leaves Black pretty much tied up for the moment. Rybka suggests the crazy 12...Rc4! - only to return to d4 after 13.Ng3 Rd4 and I'll finish this line with the verdict 'unclear' if you don't mind. Aronian's choice is much more human.

dia8A very strong manoeuvre, defending g2 and therefore practically winning the exchange with Bd3xg6. The move 12.h4, which I analysed 12 years ago as a follow-up to Shirov's suggestion of 11.f3, isn't half as effective here as it is after 10...Ba5, because after 12...d5 13.h5 d4 14.Qe1 Nxh5 15.Rxh5 dxc3 16.Bxc3 Rd6! Black has no hanging pieces now and is simply a pawn up.

12...d5! Here, of course, there is no doubt about the strength of this move.

13.Bd3 d4 14.Bxg6 hxg6 15.Bg5
As indicated in ChessVibes Openings #31, this is the critical position for the evaluation of the move 9...Rg4:

dia9White is an exchange up, but in return Black got a strong pawn on d4 and the advantage of two bishops. Before White gets the chance to finish his development, Black needs to treat the position forcefully, in order not to let his advantage vanish completely.

So, we've reached yet another unclear position in the Centre Game! Will we see it again any time soon, or is the whole variation buried again for over 10 years until another youngster gives it a try with white?

Here are a couple of general conclusions based on the material I've showed in this article. (You can find a detailed analysis of the entire game Nepomniachtchi-Aronian in ChessVibes Openings 31.)

  • Black has to make three major choices in this line: whether to retreat the bishop to a5 or to f8, whether to play d5 or d6, and where to play his rook (d4 or g6)
  • Black should only give the exchange on g6 if he can immediately active his pieces, as Aronian managed in the game
  • Although this line has a bad reputation for White, there are many ways Black can go wrong; the only clear path to advantage seems to be 9...Ba5! followed by d7-d6
  • The pawn sac 8.Qg3 is probably not enough for White if Black plays forcefully, but it sure leads to interesting play!

After all those years, I finally feel like I don't have to be ashamed anymore of my preference for this crazy opening, even if in the end it turns out to be incorrect. I hope I've shown you at least why I have been fascinated by it for such a long time - and why I tried it over and over again in so many blitz games, even against players who were so much stronger than me. Where else do you encouter such crazy rook manoeuvres, queen shuffles and opposite castling attacks after just 7 moves?
The Centre Game is a unique opening. It is more than worthy of our attention.

Arne Moll's picture
Author: Arne Moll


Signalman's picture

A great article, which I actually begin to understand !

I only recently began to play the Centre Game, and have to admit that most of my amateur opponents play an early d6, so I don't often reach the 'tabiya' with 9.a3.

On the one recent occasion that I did, Re8 was played followed up by a queenside fianchetto.

In my humble opinion not that bad but other less-active moves afterwards ( h6 , and a misguided Nh5 attacking the White Queen ) put Black in a worse position.

By the way, I learned of the Centre Game via the very good Dangerous Weapons e4/e5 book, where Andrew Greet agrees with you a lot, but gives the best line after 10. Qe3 as Ba5, stating that Bf8 is " rather passive" although he doesn't consider 11.f3, giving h3 instead, with the Flear game and then Varavin-Gusev as examples of play there.

He also quotes Nepomniachtchi as a modern advocate of the CG, but damned if I can find Greet referring to his games : he bases things around Shabalov !

Any other CG wisdom to share ?

Latest articles