Friday, June 29, 2007

How to Prepare an Opening, Part One

I am going to demonstrate a thought process that would result in choosing the main line (closed) Spanish (a.k.a. Ruy Lopez) for both white and black. The Spanish is a good choice for this demonstration, because the main line for both sides up to move 9 or so is fairly clear. It also happens to coincide (mostly) with what I play myself, so I can pretend to speak with some authority. Finally, I suspect most amateurs have at least some familiarity with the opening. This is good, because I want people to focus on the thought process behind constructing the repertoire, rather than the opening itself.

Part One is just an introduction to the process. I imagine many people reading this won't find the first 5 moves of the Spanish too mysterious. The next post in the series will be more advanced, and show how you can, with a little research, come up with your own pet lines to surprise your opponents. After that, I hope to show a real-world example of opening preparation, where an opponent's odd treatment of the Nimzo-Indian demonstrates how opening books can lead you into a dead end. Even people that have played the Nimzo for years may learn something there. So, if this post strikes you as "baby talk", bear with me. I'm hoping by the end of the series that most readers have learned a thing or two.

Let's consider the first move for White. From general opening principles, you want to control the center and open lines for your pieces to come out. To me, the most natural moves are 1.d4 and 1.e4. Other moves such as 1.c4 and 1.Nf3 are perfectly acceptable, but they don't have the clear idea behind them that 1.d4 and 1.e4 do. From my own experience, and from the opinions of many chess trainers, I would say 1.e4 is the best choice for developing players. Just consider this advice (and other opinions that follow) as a "tie-breaker". If you did choose 1.d4, the only warning I'd have is not to play some system like the Colle or Stonewall that lead to stereotyped play. In the short term they may save you time by avoiding theory, but in the long term I feel they'll hold back your chess progression.

Similar logic applies to choosing 1...e5 as black. You may find that you like other moves better. Just be sure you understand the reason behind your favorite move. For example, the Scandinavian (1...d5) may be an acceptable opening but it's not intuitively a good opening move. It violates basic opening principles, because after 2.exd5 Qxd5 your queen will be chased around, and with 2...Nf6 you are down a pawn. I now know that the Caro-Kann (1...c6) and French (1...e6) tend to result in a limited number of pawn structures that trap your pieces behind light-squared pawns, so I would steer away from those (my cautionary tale can be found in this post). However, if you understand those moves and like them, go for it.

To me, 1...e5 feels the best. It's good for the same reasons 1.e4 was: it contests the center, and opens lines for the development of the king's bishop (which will get you closer to castling) and the queen. Finally, it discourages an immediate 2.d4 by White. In any opening, if White gets two pawns at d4 and e4 without some serious concession it suggests Black was too passive.

I suspect a lot of readers are already familiar with this principle. After 1.e4, White wants to achieve 1.d4 to dominate the center, and Black wants to prevent it; after 1.d4, White wants to get 1.e4 in. Black would also like to achieve this, but it's less likely. Yet if you go over your games, I'm sure you'll find examples where after weak opening play these basic ideas are forgotten. I'm going to assume an understanding of the importance of the center in this post.

(Aside: At the time I write this, I am about two thirds the way through Marin's Beating the Open Games, which is a repertoire for Black after 1.e4 e5 and handles White's systems that avoid the Spanish before 4...a6 Ba4. This is hands-down the most instructive repertoire book I've ever encountered. It's so well written, and so objective, you could go a long way towards choosing your white repertoire by looking at what Marin considers the more difficult lines. My attitude towards some of these White options has changed from "boring", or "inferior but annoying", to "I wouldn't mind giving that a shot!" I'm looking forward to his sequel, which will handle the main-line Spanish portion of the repertoire. )

After 1.e4 e5, 2.d4 would produce the classic d4/e4 center, but it would be immediately destroyed, e.g. 2...exd4 3.Qxd4 Nc6, and so would be suspicious based on basic opening principles. So, what moves feel good here? At this point, you could see what the most popular moves at this point are to help guide you. For this demonstration I'll use Chessbase to see what the most popular moves are. References such as this, or your chess program's opening book, a one-volume book such as Nunn's Chess Openings (NCO), or whatever, are only being used as a guideline...they'll either confirm that your choice is playable, or serve as a warning that it might not be such a good idea. I made an opening book from the games of players rated 2400 or higher, to see what the "big boys" like to play. Here's the number of occurrences of the top 5 choices for White's second move:

Nf3 34110
Bc4 1054
Nc3 780
f4 565
d4 156

Again, realizing 2.Nf3 is the most popular continuation here is not earth-shattering in itself. Determining what's "popular" will be used for a couple reasons in my tutorial. Here, it's a tip-off as to what the "best" moves are and helping us to feel good about our choice. We're also going to use popularity to determine what the main lines are for our opponent so that we know which lines to learn first.

One notable exception from the above list is 2.c3. It's logical to think, "OK, I'm going to prepare d2-d4 by playing c2-c3 first". It makes sense. It can lead to a Ponziani-type opening, and was played by Morphy, but no GMs seem to play it. Let's say this was your favorite move at this point. You can either decide to make it your repertoire move, or you can try to understand why it's not popular. First, it moves another pawn instead of developing a piece, so on general principles that may not feel as good. Second, it now makes 2...d5 a decent move for Black, because after 3.exd5 Qxd5 the queen can't be harassed by 4.Nc3. In other words, c3 took away a natural developing square for the b1 N. So, you're letting Black get feisty in the center. If, after that explanation, you still like c3, then go for it...it was good enough for Morphy. If you progress to the point where your opponents are punishing you in the opening, you may decide that c3 isn't your favorite move after all.

Going through the other main moves: 2.f4 loses a pawn to 2...exd4. The King's Gambit is a real opening, but it's tactically complex and not a "common sense" opening. Odds are, if you were looking at the chessboard and knew nothing about openings, you'd consider 2.f4 a blunder (drops a pawn, exposes your king, and doesn't develop a piece). 2.Nc3 doesn't help you play d4, and doesn't get you any closer to kingside castling. Bc4 violates the guideline of developing knights before bishops. Of course, this isn't a guideline to be adhered to slavishly, but it's one that makes sense. Usually, it's clearer where you want to put your knights than your bishops, so it makes sense to delay it for a bit. Do I want to play the bishop to c4? b5? Let's see what Black does first. So, that brings us to 2.Nf3, the clear favorite. It develops a N, gets you closer to kingside castling, and supports d4.

Even if you're using a repertoire book for guidance, follow this thought process so that, at the very least, you've justified their move choice. If you think about the move for a bit, you're helping to transfer this move from short-term to long-term memory. I strongly recommend using an actual chess board, because I find it assists with this transfer.

This same technique of finding the move you like and checking it against the "book" moves is repeated until you've determined what your personal main line will be. For example, for Black we will see that by asking "if my opponent makes the most popular move, what do I want to respond with?" you can arrive at the Spanish main line of 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Bb7. I haven't actually settled on move 9 yet for my own repertoire, but Bb7 completes development (apart from moving the queen off the back rank) and is a good stopping point for these tutorials. This variation, if chosen, would be your foundation, your bedrock, your home turf, the trunk of the tree of variations. This is where you waggle your sword and proclaim, "NONE SHALL PASS!"


Just remember that it's sometimes appropriate to accept a draw as Black:



Back to our exercise in opening preparation:

2...Nc6 protects the pawn and develops a piece at the same time.

Now, if I were creating my own repertoire from scratch again, 3.d4 makes sense here. On basic principles, if pushing d4 is allowed it should be good. If Black exchanges the pawn on d4, a common pawn structure that results is called the "little center" by King and Ponzetto in their book on the Spanish, and is an example of what Kmoch calls a "jump formation":



This good for White because of the space advantage...White is less cramped and has a bit of an easier time than Black. This e4 vs. d5 center is a common feature of 1.e4 e5 openings in general, and explains White's advantage in many of the variations.

It is almost amazing how much of the theory of the Open Games can be figured out by just thinking about the center and the moves e4, d4, ...e5 and ...d5. In this case, following the basic idea of "if I can get away with d2-d4, play it!" results in the perfectly respectable Scotch opening. If you like the logic of 3.d4, and you like the pawn structures that typically arise in the Scotch, then go for it!

Back to White to play for their third move. 3.c3 would be the Ponziani opening. Theory would warn you that it's not the best choice, but it's playable. Here's a case where the guiding hand of theory might lead you to something "stronger", but if you absolutely loved it you could play it.

3.Bc4 develops the bishop to a nice diagonal and prepares castling, but it's less obvious how it addresses the center. It's not helping d2-d4, but it does help thwart d7-d5 and it also targets the sensitive f7 square. I choose 3.Bb5 because it develops the bishop and also weakens Black's control of e5. The latent threat is Bxc6 and Nxe5. This isn't possible right now (for example, 3...a6 4.Bxc6 dxc6 5.Nxe5 Qd4!), but the motif is there.

Black has many, many choices right now. Probably the most intuitive is 3...Nf6, which develops a piece, gets Black closer to castling, and attacks the e4 pawn. 3...d6 protects the e5 pawn and lets the queen bishop out. 3...Bc5 develops the bishop to a good diagonal before it can be imprisoned with d7-d6, and so on. My choice of 3...a6 is not the most intuitive, but it's the most popular. Here, rather than automatically play the "main" move, I've thought about it and have chosen it for my repertoire...I haven't had it chosen for me.

Here, challenging the bishop with ...a6 doesn't cause Black to fall behind in development (you can work this out for yourself). If White captures on c6, Black has gained the two bishops, and if White backs up (e.g. 4.Ba4) Black has turbocharged his queenside advance (...b5 will come soon). In this regard it has some resemblance to the Sicilian Najdorf. Often, wasting time in an opening by pushing a rook's pawn for no good reason is a waste of time (i.e. costs a "tempo"). Here, there are good reasons for pushing the rook's pawn. I am consciously violating a rule of thumb because there are concrete reasons to do so.

As White, I used to play the Exchange with 4.Bxc6 here. You give up the bishop pair but cripple Black's queenside majority. As someone who enjoys endgames,this sounded like a great idea, in theory... limit Black's choices, avoid theory, play positionally. Over time I found that I didn't like how my pieces coordinated, I didn't get endgames as much as I hoped, and most importantly I was missing out on the experience of playing against a variety of pawn structures. I prefer 4.Ba4 now. In both cases, I've said "this is the move that I like to play in this position", not "I don't like that book anymore, so I'm going to play something from this other book because I liked the blurb the publisher put on the back".

At this point, the main move is 4...Nf6, developing a piece, hitting the e4 pawn, and preparing to castle. I like that, so I play it. By the way, if you asked here or on the last move "is ...d5 possible?" the answer would be "not really", but you'll have to figure that out for yourself. Piket actually won a game with 4...d5 here against an IM, so it's not completely wacky, but you've been warned.

After 4...Nf6, White's e-pawn seems like it's in danger of being captured. As white, I currently play 5.Qe2 here to protect it. This is not the main move in this position, but the move makes sense (protects e4, develops the queen, vacates d1 for the king's rook) , and it's the move used in my current repertoire book. Remember: if you splurge on a repertoire book, treat them as menus of suggestions, not orders to follow. I may choose a move that the author recommends, or I may play something I like better. For example, I've gotten a lot of use out of Kaufman's repertoire book (there are very few books that provide white and black repertoires all in one volume) but I've rejected a lot of it for my own repertoire.

The standard move here is 5.0-0, and we'll see that 5.d4 is also possible. At this point there are a lot of move-order issues and unusual sidelines. In Part 2 I'll talk about how to navigate the maze that follows, and possibly help you come up with your own pet weapons to surprise an opponent.

5 comments:

wang said...

Excellent post thanks alot!

Anonymous said...

Nice post, but I would like to point out an inaccuracy. I have played the Caro-Kann for several years, and to say that it leads to 'a limited number of pawn structures that trap your pieces behind light-squared pawns' is simply false. In most lines of the Caro, black's light-squared bishop is developed quickly and often traded early. The Classical variation (4...Bf5) is the best example. As a model I would suggest one of the Polgar-Bareev games from the recent FIDE candidates matches.

Unknown said...

i'm gonna print out your posts, bind them in a book, and sell the book for 60 bucks a pop, and i know everyone who plays chess in the whole world will buy the book because it's underpriced....

Grandpatzer said...

anonymous: it's true that the bishop isn't trapped in the Caro-Slav complex (as opposed to QGD/French). I'm thinking of the positions in general. I have played the Slav, and looked at the Caro a bit, and a typical pawn structure for both is pawns on c6 and e6, with Black's big pawn breaks being ...c5 and ...e5.

To me, Black is often crouching behind the same pawn structure in both the Caro and the Slav, and my impression is that people that suggest a repertoire using both the Caro and the Slav do so specifically to get this similar pawn structure from both openings. That's what's limiting to me.

Grandpatzer said...

edit: for some insight into the Caro-Slav pawn structure(s), see Pawn Structure Chess by Soltis, Chapters 1 and 2.