Preparations of an amateur: King's Indian without c4

As a mediocre club player, my preparations for a game is usually very poor. This time, I thought I should take it more seriously and document my preparations against the King's Indian, fianchetto without c4 (A49) - which I will most likely meet on Wednesday.

I play in an OTB tournament, round robin, where I am rated 10 out of 12 - and my statistics so far show that I am the worst player. That is ok - I play against fairly good players and have held my own in most of the games before difference in class have shown in understanding of a position or pure lack of mistakes. So I have learnt a few things.

On Wednesday, I will play Black against a player who seems to play the King's Indian with a g3/Bg2 setup when given the opportunity - which seems a bit slow:

I say slow because g3/Bg2 is solid, but not immediately threathening for Black.

So, I don't know too much about the A49 ECO code, and I am not going to buy any books about it. A quick web search brings me to a few databases, but nothing that analyses the opening in normal language. Youtube doesn't seem to have anything that helps, so I am left to myself. Which is not a bad thing! 

Besides, I am not all alone. Chessbase has a great tool to look at openings and find lines that might be possible to use. database.chessbase.com lists millions of games, with stats. They have app. 20.000 games with the position above.

So, what should I play here? Of course, d5 looks natural. It establishes Blacks control of e4, blocks the long light diagonal and opens d7 for my bishop or my knight. It is the most used move in the database, with a 56 % score for White - not catastrophic for Black, but is there anything else?

3...b6 would create the possibility of a clash over the long diagonal; meeting 4. Bg2 with Bb7. This is also the 2nd most used move here, with a 57 % score for White; marginally worse than 3...d5.

3...Bb4+ is a move you would often see at my level; striking out immediately, even with a check! However, it is rarely played by really good players - and has a terrible score: 76 % for White - as the early strike is certainly premature and not preemptive:

So, is there a better and as aggressive move for Black here? Of course. It's the third most popular choice, with a slightly better stats for White than Black (53 % vs 47 %) - and it puts pressure on the d4 pawn:

With c5!? it seems to me that Black is being a lot more direct in his play - it would be logical to follow up with cxd4 - removing a center pawn for White.

However, the very best score for Black in this position is 3...b5!? This is played by high rated players, so it must be good, right? Also, the score for White is only 45 % - so I must play b5!

But - this is where I hope I have changed as a player over the last year: Previously I would just find a line with good stats, memorize the first 5-6 moves and hope to end up here, for instance:

The 3...b5 is an important move in the Benko gambit (which I don't play) - so if you are a Benko player, you may feel at home here. I don't. Besides: After 3...b5 White scores great with 4. e4 Nxe4 5. Bxb5 Bb7 6. 0-0 Be7, where White are clearly ahead in development.

So - 3...c5 it will be. But what then?

It is impossible to go over all deviations. I will trust my thought process at the board if I have read my opponent wrongly in my preparations. Even allowing for only two options on each half move from here will give you a headache by move 6. 

So, judging by very few previous games of my opponent, my guess is that he will play his solid and slow game - 4. Bg2 (55 % for White); the other main move is 4. c4 - but that would possibly transpose away from the setup he likes; though c4 would probably be played by White later on anyway.

After 4. Bg2 it is simple: I take his d4 pawn, both as it is a lot more active than he seems to prefer - and the stats are clearly better (50 %) than the obvious alternatives: d5 and Nc6.

Then White will most likely recapture, though there seems to be time to castle first. That move order is not very important, it seems - we will most likely go like this:

Again, judging by the previous games, I would expect White to play Nf3 here - it brings the knight back as a defender of his own king; though Nb3/Nd2-Nf3 will do the same.

Either way, I will try to push the d-pawn to d4, and support it with Nc6. If White plays 6. c4 (which is by far the most used move), we might end up like this after 9. e3:

So, nine moves in - I have interesting pawns in the centre; it is unclear how strong they are, but they must certainly pose some annoyance for White, at least. They can definitely not be ignored, so there is no time for White to for instance fianchetto also the other bishop, which I noted my opponent did in a previous game.

So, 9...Bc5 is most used and has good stats (48 % for White). But why should I accept an isolated d pawn? 

Jeremy Silman has written several articles on isolated pawns here on chess.com, and I must admit it is one of the concepts that I certainly understand the least. But, to borrow Silman's words: 

"If you are going to enter these kinds of positions, you need to adjust your mindset to the “It’s strong!” view rather than your negative “I have to trade the sucker off fast!” association." (See The Art and Science of the Isolated d pawn, chess.com)

This pawn looks strong; and if I follow a couple of unlikely lines here, it is possible to start to think I know why.

Kasparov ended up in this position once, and seems to have used the pawn in order to crush his less esteemed opponent:

This is a more likely variation; not taking the pawn, but activating the rest of the White army instead:

So, will these preparations prevent me from making the same mistake as in my last game, when in a completely winning position I ignored the 14 moves that would win, and chose one of the two that would definitely lose? Of course not. But when I meet the King's Indian, fianchetto without c4 - I will be slightly better prepared. Bring on the game.

Correction: In one of the diagrams, I say the opening transposes to a normal Queen's gambit, while it actually transpose to a Closed Catalan (E06).