I would like to present you with an excerpt from one of the last emails that I sent to my Inner Circle. In case you would like to receive this kind of emails, please use the friendly yellow form on the right to subscribe.

Enter the Caro-Kann. If I have to explain why, I’d say that it has to do with the inherent solidity of the structures. The Berlin is also solid (I have also prepared that one at depth!) but the Caro feels more compact. After all, the pawn doesn’t go that far, only 1 square ahead!

When studying Fischer’s games I noticed that he played all the lines against the Caro-Kann (except the Advance, which at that time was considered completely harmless). Compare that to his almost exclusive use of the Sozin against all Sicilians. My impression was that he wasn’t sure which line was best against it, so he kept varying. In the last 5 Caro-Kanns he played, all in 1970, he played 3 ed in game 1 against Petrosian in the Match of the Century, 3 Nc3 in game 3, and three times he played 2 d3 (against Marovic, Hort and Hubner).

Curiously enough, the same situation of almost all lines being more or less equally playable applies to today. But here’s the thing that attracted me to the idea of the Caro – almost all of them can be easily solved.

Let’s start one by one. The Panov is practically sterile after 5…Nc6 (after 3 ed cd 4 c4 Nf6 5 Nc3) even though the other lines 5…e6 and 5…g6 are perfectly playable too. The Exchange Variation after 3 ed cd 4 Bd3, successfully used by Fischer in that famous game 1 of the Match of the Century, is rather toothless nowadays. Black has nothing to fear there after playing logically 4…Nc6 5 c3 Nf6 6 Bf4 Bg4 7 Qb3 Qc8. In the reverse Carlsbad structure with the bishop out on g4 the position is as safe and solid as possible.

2 d3 can be met either by the fianchetto (2…d5 3 Nd2 g6 4 Ngf3 Bg7 5 g3 e5) or the central development of the bishop after 2…d5 3 Nd2 e5 4 Ngf3 Bd6. I would go for the former, in order to avoid 5 d4 in case of the latter, but it is largely a matter of taste.

The Two Knights, an early Fischer favourite, can be met in various ways. My preference would be the established 3…Bg4 4 h3 Bf3 5 Qf3 e6. What can possibly go wrong there?

And so we arrive at the main lines. Nowadays there are two critical ways to tackle the Caro – the Advance Variation (in particular Short’s plan with 4 Nf3 e6 5 Be2) and the Main Lines with 3 Nc3 de 4 Ne4.

It is actually the Advance Variation that bothered me when thinking about the Caro-Kann for Black. It looks unassuming and not really threatening, but that space advantage White has becomes annoying as the game goes on (I suffered some unplesant losses in several training games I played). And also there was the problem of choice of lines. There are so many plans and move orders to choose from and I found them rather confusing because I couldn’t see a clear-cut plan behind them.

And then an obscure line appeared. It was the absurd-looking check 5…Bb4 (Black can also play 5…Nd7 first and then 6 0-0 Bb4). It was a game by Carlsen that drew my attention to it – he used this line to beat Giri in a rapid game last year. True, he also used this line to famously lose to Sjugirov in 2010, but then I wasn’t paying attention.

The reason why I became attracted to the idea is that here Black’s play had purpose – the bishop will drop to c7 to attack e5 after …f6, or to b6 to attack d4 after …c5. On the kingside there was also a clear way how to place the pieces – Black plays …h6 to secure the Bf5, plays …Ne7 and castles. Then he tries to achieve …c5 and/or …f6. And that’s it! Very simple and straight-forward. The line is becoming popular as after Carlsen it was Mamedyarov and Andreikin who continued to use it on a regular basis. There is, however, some hidden danger in these lines when White plays Nh4 and Nh5 and for now this seems to be the best way to play for White.

The main lines after 3 Nc3 de 4 Ne4 Bf5 (I was never really attracted to 4…Nd7 or 4…Nf6) 5 Ng3 Bg6 also looked uncomfortable. If the sidelines like 6 Nh3 or 6 Bc4 can be dealt with, the main line after 6 h4 h6 7 Nf3 looked problematic. I always had the hunch that here 7…e6 should be played, instead of the more popular 7…Nd7. Not because it is the better move, but because it is somewhat less common and gives Black some additional options later on, such as playing …c5 and …Nc6. Again it was Carlsen who employed this first at a top level – he used it in the first black game of his match with Anand in 2013. Refinement came several years later, when Black learned not to give the check on b4 on move 11 and play 11…Be7 immediately – the point being the pawn on d4 hangs now.

The crucial games were played by Giri and Mamedyarov:  Wei Yi-Giri from this year’s Wijk and Saric-Mamedyarov from Baku’s Olympiad. Both of them suggest that Black is in excellent shape!

You can download the lines mentioned above here. They are far from conclusive, but they can serve as a good starting point for creating a Caro-Kann repertoire. Who knows, maybe soon enough I will use them myself!

But there is also Part II of this Repertoire and it can come directly to your Inbox on Saturday if you join my Inner Circle before then!

Before I present my analysis, a short round-up of my Llucmajor tournament. After not playing for some time I was wary that my head-to-hand coordination might suffer. The blitz tournament before the main event served me a very good purpose to get some practice. I made a conscious decision to trust my feelings more and, crucially, not always try to validate those feelings with calculations. This was necessary because I knew that my calculations wouldn’t be on the level of some time ago.

And things went so well! I was playing quickly and confidently. Topalov’s words that it is not always necessary to play the best move resonated well – I knew that my understanding is good enough and that I will never make a bad positional move. As for the verification by calculation, I tried to go for width over depth – check many more moves in the starting position but only calculate them 2-3 moves deep. This is a very good method that insures against blunders.

I had a great run and had 6/8 before the last round (and had I won a completely winning queen endgame against IM Kohlweyer in Round 6 it would have possibly been more). The game I won in Round 8 actually served as an inspiration for this post. My opponent played the Schara-Hennig Gambit and I quickly got a position that was close to winning. I didn’t win quickly, but my technique to convert the advantage of the bishop pair in an open position wasn’t bad at all. You can see this and the other games in the game analysis below and you can download them here.

The games in Llucmajor started at 8.30pm and finished well after midnight. That was OK, but the last round started at 9.30am and that turned out to be a big problem for me. After I finished my game in Round 8 I waited for the pairings until 2am when I discovered that I was to play 4th seed Daniel Fridman, rated 2605. He plays 1 d4, 1 c4 and 1 Nf3, so my preparations lasted until 4am (and I had to limit them!) Then I couldn’t fall asleep for the remaining of the night and got up at around 8am. I was feeling horrible and even though I managed to get my preparation in and obtain an easily-drawn endgame, I couldn’t calculate the simplest lines. Soon enough I blundered a pawn and that was it. It was certainly a disappointment, a good tournament spoilt at the end. But there are positives as well, the newly-found playing algorithm looks very promising.

Now onto the refutation. I would like to repeat here what I said when I talked about the refutation of the Budapest Gambit – by refutation I don’t mean a forced win, not at all. What I have in mind is that “White obtains a stable and safe advantage out of the opening in all the lines and at the same time Black’s position lacks perspective and has no way to evolve.” In the case of the Tarrasch Defence the positions that arise after best play lead to an endgame where White has the advantage of the bishop pair and Black’s only prospect is of a long and arduous defence. If in addition to the objective evaluation as “very unpleasant for Black” we add the psychological factor that Tarrasch players like active play and thrive in the IQP positions, then we can safely conclude that the variation with 6 dc is really a “refutation.”

You can download the analysis plus three of the games I mentioned in the comments here.

As you can easily see on the right side of this text I have a subsription form where I invite you to join my Inner Circle. You may be wondering what this is actually about, so I would like to share with you a sample of my newsletter. I write these on a weekly basis (if I have the time even more often) and they are all original pieces of writing on various topics – some are purely chess-related (like the one below), others have to do with chess psychology, others still with other aspects of our game (I recently wrote about cold showers. I leave you to figure out what does that have to do with chess.) So if you would like to receive emails of this kind (and also some welcome gifts when you subscribe) use the form on the right and join the Circle!

Without further ado, here are the two ideas against the King’s Indian Attack:

The King’s Indian Attack. Ever since I was a child I had mixed feelings about it. The main reason – because there was never an attack. Of course, I saw Fischer’s famous games where he was destroying his opponents (Myagmarsuren is the most famous destruction), but when I would analyse a bit on my own I could never crash through Black’s kingside, especially if he played …h6 when White threatened to play h6 himself (for some reason they never played this against Fischer, always allowing him to push h6 and fatally weaken Black’s kingside).

When I had a chance to play the Attack myself (this was in 1995), guess what they played against me? …h6 of course, leaving me fuming and cursing myself. But somehow I managed to push g4, threatening g5 and also c4 and I won the game, but never dared to try the Attack again.

Then came times when the Attack was played against me. It is usually played when Black had already committed the pawn to e6, in my case after 1 e4 c5 2 Nf3 e6 3 d3, but then I was happy to use the harmonious development with 3…Nc6 4 g3 g6, followed by …Bg7,…Nge7 and …d6. It is still one of the best systems against it.

The Attack has recently had an increased importance because a lot of players use 1 Nf3 as a way to enter the Reti or some other opening, depending on Black’s choice. And many Black players rely on the QGD, so we often have something like 1 Nf3 d5 2 g3 Nf6 3 Bg2 e6 4 0-0 Be7 and now White plays 5 d3, not wanting to go into Catalan waters with 5 d4 or Reti with 5 c4 or 5 b3. So Black continues 5…0-0 6 Nbd2 and we are on the verge of the King’s Indian Attack.

Looking at it from Black’s perspective some time ago I was attracted by Aronian’s choice of 6…Nc6, with the idea of 7 e4 de 8 de e5! and there is a transformation in the position which is quite pleasant for Black. This is a good idea for Black, especially important is that it appears on the board early and it is Black who dictates the changes in the position!

And then Kramnik came. His idea (not actually his, it has been played before but it was largely forgotten) is deeper, but at the same time it is also universal. Black pushes the pawn to a3, provoking b3 (Fischer didn’t allow this – he played a3 himself when Black threatened to push …a3. But Kramnik’s idea works in this case as well – it is only important to keep the bishop on c8, to have the pawn on e6 protected). Then, when White reaches h5 with his pawn, he plays …f5, stopping White’s kingside play in its tracks. Here’s the position after 13…f5! from his game against Nepomniachtchi from the opening blitz in Zurich, some 2 weeks ago:

If White doesn’t take en passant then he has nothing. If he takes, he is worse. Nepomniachtchi took and lost 12 moves later. It is worth noting that Kramnik’s idea can be implemented also with a knight on e8 instead of d7, from where the knight can go to c7 to bolster e6.

So there you go. In case you needed a reliable weapon against the misleadingly-named King’s Indian Attack – now you have two!

In my previous post I wrote that I will take a look at my games with black in the PRO Chess League. They were notable because the variation I used exceeded all my expectations and the positions I got very excellent.

It was one of those ideas that I have briefly looked at but never analysed properly. Since I decided that I would play the Nimzo against 1 d4 my choice was the …g6 idea in the QID after 1 d4 Nf6 2 c4 e6 3 Nf3 b6 4 g3 Bb7 5 Bg2 g6!? The good thing about this idea is that it can be universally used after 1 Nf3 as well, with a transposition very likely. I dubbed it The Double Fianchetto Solution and I wrote about it in more detail in my newsletter (if you’re interested, please sign up using the form on the right). That is what happened in my first game with it:

Not a bad start at all! Both the opening and the result were very satisfactory.

The second game was the biggest test. I played GM Vidit, rated 2689 at the time, who served as Anish Giri’s second in the Wijk aan Zee that finished just days before our game. I could expect some top-notch preparation and he didn’t disappoint.

A very lucky win for me and the opening could have gone better. As Capablanca used to say, one should play the openings that bring good results, irrelevant of the positions you get from them! And since black’s opening could be improved upon, I decided to continue to use the system.

I missed my chances, but the opening was a breeze again. The next two games can actually serve as a completion of the repertoire based on the double fianchetto as they showed how black can use the 2…b6 move order against various systems that white can play after 2 Nf3. I discovered this in the mid 00s when my repertoire was based on the Nimzo and the QID – playing 1 d4 Nf6 2 Nf3 b6 gives black additional options agaist the London and Colle Systems because he can use the double fianchetto and obtain more dynamic play.

I felt under some pressure after the Grunfeld-like choice of 7…d5 (now I would prefer 7…d6) but there was never a clear way for a white advantage. And I won again!

The last game was a worse version than the London System as white played 3 Bg5, giving black a tempo with …Ne4.

Again smooth sailing for black! And a final 4.5/5 against such strong opposition is definitely an excellent advertisement for the variation. The practical point of playing it (it can be used against both 1 d4 and 1 Nf3) was the decisive factor in my choosing it. Another important aspect was the psychological – I didn’t mind getting equal and simple positions with several pieces exchanged, thus risking a draw against lower-rated players. The reason was that in rapid games mistakes happen in all positions so a chance to win the game will present itself even in the dullest of positions. Now that I think of it, perhaps this is applicable for classical chess as well…

Of course, if one wants to complete the repertoire then the Nimzo, the Catalan and the English 1 c4 need to be taken into consideration, but nowadays the trend is to avoid the Nimzo, so I expected people to allow me to play the line in this tournament. Personally I still feel that I need to prepare the line more deeply if I want to play it in long games, but it certainly showed great potential. For now, let’s keep my future opponents guessing!