Transposition Gone Wrong: Attempting to reach the Benko Gambit

Happy New Year! May 2018 bring us the best chess... Also, fame and fortune.

I love playing the Benko Gambit as Black, so I try to reach it whenever I can. But it doesn't always go so smoothly. Normally the Benko Gambit is reached after 1. d4 Nf6 2. c4 c5 3. d5 b5:

Typically White will play cxb5 at some point, and Black will respond with a6 (a pawn gambit), and White can choose to take the pawn with bxa6 (gambit accepted--the most common response) or to push with b6 (gambit declined). When the gambit is accepted, Black will fianchetto the king's bishop, put the knight on f6, play d6 (some players prefer e6), castle kingside; White will typically get e2-e4 in (even though this is not an e4 opening). A position like this might arise:

Black will eventually recapture the a6 pawn, and White will have a passed a-pawn, but Black will have tremendous activity on the queenside, with moves like Bxa6, Nbd7, Qb6, Rfb8 (after Black castles kingside), etc. Note that in the "old main line", which I don't particularly fancy, Black plays Bxa6 before White plays e4, so that when White plays e4, Black responds with Bxf1, and White plays Kxf1 and will manually castle their king with g3 and Kg2.

I've had quite a bit of success playing the Benko as Black. This graph shows the win percentage (in green), draw percentage (in blue), and loss percentage (in red) by opening when I play a rapid or classical game (> 8 minutes) as Black on lichess.org. In fact, my highest win percentage comes from games where I've played the Benko (10 games, win percentage = 80%; see middle of the graph):

So I'm somewhat justified in trying to reach this opening when I can. Against 1. e4, there are a few ways to try to reach this opening. An uncommon one is the St. George defense, with 1. e4 a6 (since both e4 for White and a6 for Black are moves in Benko lines). A more common opening that allows one to hope to reach a Benko position is the Robtasch (or "Modern") defense with 1. e4 g6, which is what I played in this game (incidentally, notice that the Robatsch seems to be one of my worst openings in the graph above). In this game I was facing a stronger opponent (2142 blitz on chess.com), and we reached the following position after 1. e4 g6 2. d4 Bg7 3. Nf3 a6 4. c4 d6 5. Nc3 Nd7 6. Be2 c5:

Up to here I've played all moves that are consistent with the Benko so that I can transpose into it were the opportunity arise. And we're really close to it: If White plays d5, I'll respond with b5 and we will have reached a position consistent with the Benko gambit. But my opponent doesn't have to humor me, and indeed he chose not to, and simply played 7. O-O. But I was not ready to give up though I should have been, and I played 7. ... b5? 8. cxb5:

And here I played a move that would simply be impossible to play in the Benko gambit since White would have played d5 earlier. I played the ridiculous-looking 8. ... cxd4?! 9. Nxd4 axb5?? (9. ... a5) 10. Bxb5:

And once the smoke has cleared, this is not a Benko gambit, and White has two connected passed pawns on the queenside! (With a normal Benko gambit White would only have a passed a-pawn.) Stockfish gives White a two-pawn advantage at this point, and my opponent went on to kick my butt. I later gave up one of my knights for his two pawns but it was already too late.

So, lesson learned: Don't part with all three of the c-, b-, and a-pawns in the Benko Gambit, because White will have two connected passed pawns on the queenside. It seems obvious, but I had to learn the hard way.

The full game is available here:

Positional chess: an exact science?

In most books on chess strategy, authors explain why such and such a move was the best in the given position, or even, they ask you to find the best move in a position. Some do this by giving a multiple-choice question, where you have to choose the best plan (e.g., Alfonso Romero in Ultimate Chess Strategy Book); others give a diagram and simply ask you to find the best move (this is the case for Aagaard in Excelling at Positional Chess, which was the subject of an earlier article).

The underlying postulate is that given any (or most?) position exists an objective best move or plan, and using abstract thinking you are asked to find it.

I used to tell one of my friends – who doesn’t like chess at all so needless to say that she didn’t care – that one of the big differences between the middlegame and the endgame is that very often in the endgame, one and only one move wins (or draws), and all other moves lose. Whereas in the middlegame, you can play many moves that are just as good as the others: it’s much less exact.

So what’s the answer to that question anyway? Is positional chess an exact science, much like the endgame, as Aagaard sees it? Is it a multiple-choice decision, where some plans are good but not equally, so they are not awarded the same amount of points by the author when you choose them? Or are there some moves which are equally right?

I don’t have the definite answer to that question, but I recently ran into a book which brings the most interesting (and nuanced) answer so far. The book is written by GM Lars Bo Hansen and is called Foundations of Chess Strategy. I haven’t read the whole book – I actually found it on GambitBooks.com, looking at their list of books, which comes with a sample of each book in PDF format; the sample for Foundations of Chess Strategy is here.

The author says that the best move in one’s position depends on their playing style. In the chapter “The Opponents: The Role of the Human Factor in Chess”, Hansen describes his viewpoint as thus:

The players in a chess game are humans [...], and the choices they make are influenced by their background, experience, self-confidence, personality, etc. This means that what is the right choice in a given position for one player is not the right choice for another player with a completely different personality and chess style. Therefore there is no ‘best’ choice in a (strategic) position – no ‘one size fits all’ approach! The right choice of plan in a given strategic position should not only be determined by purely chess reasons. [...] This means that we should give up the assumption that in a given strategic position there is one best way to play which should be chosen by any player in the given position against any opponent sitting on the other side of the board.

Hansen appropriately gives the example of a positional (“solid”) player versus an attacking (“sharp”) player. When discussing a line chosen in the Queen’s Gambit by Karpov (solid) and Kasparov (sharp) – who did not choose the same line –, Hansen writes: “according to the statistics in my database Karpov has an almost 80% score against top-level opposition in ‘his’ line, while Kasparov even displays 90% in ‘his’! Despite their magnificent talents I doubt they would achieve the same scores if we switched the variations [i.e., Karpov playing Kasparov’s line and vice-versa]. Obviously they would still score well, but I don’t believe it would be this high.”

The author does make an interesting point. As he accurately points out, we all have played in positions in which we knew we had an objective acceptable position, but we still didn’t feel comfortable playing it. I might even add that although opening theory is huge and provides us with a wide choice of openings, only a few truly fit our style.

I think the example of Karpov and Kasparov speaks for itself. But don’t get me wrong: I am not discrediting neither Aagaard’s nor Romero’s approach. I have Aagaard’s book, and I think it’s excellent, and it has helped me a great deal. I think that all in all, critical judgement is still essential, even when reading GM books (although let’s face it, they’re more right than I am hehe). More importantly, you should choose moves (and openings for that matter) that truly fit your style, and that give rise to positions in which you are comfortable. Don’t choose a line because your chess idol or favourite author (in his positional exercises!) praises it.

Movement-limitation patterns in pieces

Arrangement between two pieces (or more) can result in limitation of movement for one of the pieces, or both of them in mutual-movement limitation. That's well known. The typical example is the Bishop and Knight separated by two ranks or files (see Diagram below), where the Knight sees all of its movements stopped in the direction of the Bishop. As for the Queen, being on the adjacent square leads to the same result -- the limitation of the Knight's movements towards the Queen.

One less well known pattern is the one involving the Rook and Knight. Perhaps you know that a Rook and Knight on two adjacent squares on the same diagonal leads to the limitation of the Knight's movements in the specific direction of the Rook (i.e., it blocks two squares; see in the Diagram below and imagine the Knight one step closer to the Rook). But there is another important pattern: when the Rook and Knight have a square between them on the same diagonal. In that case, the Knight sees four of its available squares taken away, with devastating results when the Knight is on the side of the board. When in the corner, this is the way to trap the Knight with a Rook (see Diagram below).

I remember using this pattern during a middlegame, where my opponent's Knight was on the side of the board, and my using this pattern led to the trapping of the Knight (and eventually to its capture). You might also be interested in the game Karpov-Ftacnik, Olympiads 1988 (you can search on http://www.chesslive.de/), which involves a pawnless Rook vs Knight ending.

-

The last pattern I will discuss here is of another kind. It shows the cooperation between two Knights separated by two ranks or files, given that they are on the same file or rank (see Diagram below). In that case, both dark- and light-squares in-between them are under their control, except for the squares directly aligned with the two Knights.

-

In the game Harikrishna-Volokitin, 2005a, White could have used this pattern to win a pawn:

1. Bxf5! Bxf5 2. Rf4 Kg5 3. g3!!

-

This move protects both the Rook and Knight, but most importantly now, whatever Black plays, White's next move is to take back the Bishop with the Rook and then throw the powerful Ne4+, winning back the material. A possible continuation is 4. ... Ne6 (attacking White's Rook and defending Black's own Rook) 5. Rxf5+ Rxf5 6. Ne4+ (see Diagram below).

And we see the pattern. Black needs to move his King, but due to the configuration of the Knights, it can't go anywhere to protect the Rook, and White ends up being a pawn-up in a two-knights-vs-two-knights ending.

Exchange sacrifice

In his book Rethinking the Chess Pieces, GM Andrew Soltis gives the following guidelines to know when an exchange sacrifice works:

• when it changes the impact of the minor pieces (e.g., taking a bishop to enhance the power of your own bishop of that color);
• when rooks can't act like rooks (i.e., blocked files);
• when there are many minor pieces left;
• when it destroys the opponent's pawn structure;
• when you thus get the bishop pair.

The following example is taken from Ridli-Volokitin, 2004, and can be found in Volokitin's excellent Perfect Your Chess, which will be discussed in another article, along with Soltis' book mentioned above. The following position arose after 15. axb3 Ra2:

White responded with 16. Ba1, when Black threw the excellent positional sacrifice 16. ... Rxa1! 17. Rxa1 Bxd4

This not only technically loses the actual value of only one pawn, but mostly, it prepares to shatter White's pawn structure and completely block in White's other bishop. The rook being attacked, White had to move it, giving Black the opportunity to take on e3, which was the key to the whole combination.

18. Ra8?! Volokitin suggests Ra3, to protect the b3 pawn; 18. ... Bxe3! 19. fxe3 Qxd1 20. Rxd1 Nxb3.



White's hemmed-in bishop proved to be a problem, and Black went on to win the game.