Monday, July 28, 2008

A Thought on GM Openings

I just renewed my lapsed subscription to New In Chess, and I've been going over some of the recent tournaments. My repertoire database in Chessbase is starting to get to a decent size, so for kicks I tested it out by generating a repertoire using the games from the Corus tournament (Section A). For those not familiar with this feature: Chessbase can search a collection of games and compare it to a database of lines you consider to be part of your repertoire. It then spits out all games that match your repertoire. This search actually did produce some significant games (Carlsen-Polgar and Radjabov-Anand, for example).

In the course of reviewing some of these games, I was struck once again by how many of these GM lines seem to teeter on the brink of disaster. Some lines are just positionally awkward, and can only be justified tactically.

I would compare these cutting-edge openings to cutting-edge aircraft. Planes are now being designed with "negative stability". These aircraft are inherently unstable and require a "fly-by-wire", computer-controlled system in order to stay aloft. If a system were to fail, the plane would be uncontrollable. In contrast, a plane designed with "positive stability" will tend to maintain its attitude without external control. The pilot can't just doze off, but the plane has a natural tendency to stay aloft.

My preference is middle of the road...major lines that you see high-rated players play, yet fairly well grounded in general chess principles. No Colle or King's Indian Attack for me (though those would fit the bill as openings with "positive stability"), but no Smith-Morra Gambit or Botvinnik Semi-Slav/Anti-Moscow either.

To each his own. There's an argument for playing highly tactical openings, in that they force you to hone your ability to calculate. For some of these GM lines, though, I think the average club player would just fall into the trap of spending time memorizing lines rather than understand why certain crazy moves are essential.

Saturday, July 12, 2008

A Gem of an Endgame Book

Back at my first chess club we had a National Master, originally from Russia, that gave group lessons. He would set up positions for us to play against him, and then we would discuss what was played and what proper play would have been. Many of these lessons were on endgame technique. At least one group member questioned the value of the lessons, along the lines of: "What are the odds that I'll see this position in my own games? Maybe I should just work on my openings." I, however, felt I really "got" what the instructor was doing and why.

I have been working through Chess Endgame Training by Bernd Rosen, and I'm strongly reminded of these group sessions from years ago. This endgame book, while instructive on its own, is actually structured pedagogically so that an instructor can use the material in their own group lessons. The bulk of the book consists of the positions and their analysis, with the final section of the book giving advice on how to teach the material.

I am about halfway through this book, and I am completely enthralled by it. This review from struck me as being relatively accurate, but lukewarm. In particular, this book is not intended to be the first endgame book you ever read. However, I would say that players class C and above that have already spent time covering endgame basics can get a lot out of this book.

First, let's look at the breakdown of chapters:

  • 6 chapters on pawn endgames
  • 1 chapter each on knight, bishop, and bishop vs. knight endgames
  • 4 chapters on rook endgames
  • 2 chapters on queen endings
  • 1 chapter to test your knowledge

plus the final chapter on Advice for Teachers. This may seem heavily weighed towards pawn and rook endgames, but I think that this is justified. Pawn endgames serve as a foundation for other endgames (for example, you always have to consider possible transitions from an unclear piece endgame to a concrete pawn endgame). I also think that pawn endgames are more coachable than the others, and that they are good tools for practicing visualization and the calculation of variations. Rook endgames are the most common endgame in practice, and are technically demanding, so emphasis there is warranted.

Each chapter contains roughly 8 easier problems and 8 tougher problems. What makes this book special for me is how the more advanced problems are related to the easier. The author is making connections between the classic endgame positions that you may have seen time and again to more complex positions that resemble actual games. Someone casually flipping through the book in a book store may think that this is old material. For example, many of the positions in the first two pawn chapters are old chestnuts. However, dig deeper and you'll see the author bringing in fresh material (fresh to me, at least) that extrapolate from these basic concepts.

The end result is an endgame book that, in a relatively concise manner, manages to both cover the essential basics of endgames and provide examples to challenge good endgame players. I would estimate that players from Class C to Class A can benefit from this book. Weaker endgame players will find many of the endings difficult, and stronger ones will find much of the beginner material obvious, but there is a lot of middle ground here. The easier positions demonstrate standard endgame techniques, but many of the harder positions demonstrate that sometimes there's no substitute for brute-force calculation.

I'm going to share a handful of the many endgames from Chess Endgame Training that, to me, shone like jewels. I've struggled with what to include here, since there is so much to enjoy. The following are taken from the first four chapters. I'm going to show a basic endgame, followed by a more advanced endgame of the same theme. In some cases I've truncated the analysis--in part for brevity, and in part because I want to encourage people to check out Rosen's book themselves. I've also added my own comments and analysis in places. Once again, I'm trying to cleave closely to the Nunn Convention for annotating endgames, so you will see a lot of "!"...there's not much room for error in these endgames.

Chapter 1 covers the square of the pawn, the opposition, waiting moves, and key squares. For this introductory chapter, all of the examples look like standard positions you'd find in endgame manuals, so I've selected a pair that demonstrate the opposition.

Black to move

This is a standard position on opposition. Its position in the chapter (1.6) places it on the more difficult side of the "easy" half. I think that this is a good test to see if Rosen's book may be for you. If this position elicits a "yeah, yeah, I know how this position is played, with either side to move, and can do it in my sleep", then you're ready for this book.

(The solution is 1...Kd8! taking the distant opposition, e.g. 2.Kd5 Kd7! 3.Ke5 Ke7!).

Kranki-W. Lange, Bad Oeynhausen 1940
Black to move

Although this position looks like a study, it's taken from an actual game. This is a great example of how the opposition is used to invade and capture your opponent's pawns. It's pretty obvious that Black needs to seize the opposition, but less obvious to see how they can penetrate. If you still don't have a feel for how to use the opposition, this could be a good position to play against the computer.

1... Kg2! seizes the opposition 2. Kc3 (2. Kc1 was played by Fritz vs. me: ...Kf3! 3. Kd2 Kf2! 4. Kc2 Ke2! 5. Kc3 Kd1! 6. Kd3 Kc1! 7. Kc3 Kb1! 8. Kd3 Kb2! 9. Kd2 Kxb3!) 2... Kf1! the only move that doesn't cede the opposition 3. Kd2 (3. Kd3 Ke1! 4. Ke3 Kd1!
5. Kd3 Kc1! 6. Kc3 Kb1! Note the theme of "scraping off" the opponent's king on a barrier such as a pawn or an attacked square, causing them to lose the opposition) 3... Kf2! 4. Kd3 Ke1! 5. Ke4 Kd2 6. Kd5 Kc3 7. Kc6 Kxb3! 0-1

The next three positions from Chapter 2 demonstrate body-checks and the opposition. The first is a standard position you'll find in many books:

Schlage-Ahues, Berlin 1921
White to move

1. Ke6! Kc3 2. Kd5! the body-check (in the actual game, 2. Kd6? led to an instructive draw: Kd4! 3. Kc6 Ke5! 4. Kb7 Kd6! 5. Kxa7 Kc7!= A standard technique in K+RP vs. K endgames) 2... Kb4 3. Kc6! Ka5 4. Kb7! Kb5 5. Kxa7! Kc6 6. Kb8!+-

However, the next two are far from obvious:

White to move

Seizing the opposition with 1.Kd5? draws, as does 1.Ke5? and 1.f4?. However, 1.Kd4! wins by force (see Rosen for details, or play this against a chess engine).

White to move

A common theme in pawn endgames is defending from the rear. For example, 1. Kc3? Kg4 2. Kd4 Kf4! 3. Kxd5 Ke3!=. White has to take a more coy approach to winning the pawn:
1. Kb4! Kg4 2. Kc5! Kf4 3. Kd4!+-

I feel these last two endgames demonstrate a point: sometimes there's no substitute for calculation. Some people prefer books that help them "understand the ideas" behind endgames, openings and whatnot, and certainly there are standard endgame positions where such commentary is helpful. On the other hand, there are endgames where rules of thumb and pattern recognition are actually detrimental. For example, in 2.10 using intuition and taking the opposition is wrong.

The next two examples from Chapter 3 cover the protected passed pawn:

White to move

1. h6! Kf8 2. g5! creates a protected passer 2...Kg8 (2... f5 3. gxf6! Kf7 4. h7!+-) 3. Kd2! (3. Kc2? The king cannot dilly-dally; the passed pawn doesn't win on its own: 3...Kh7! 4. Kd3 f6=) 3... Kh7 4. Ke3! f6 5. gxf6! Kxh6 6. Kf4! g5+ 7. Kf5 Kh7 (7... g4 8. Ke6! g3 9. f7! g2
10. f8=Q+!+-) 8. Kxg5 Kh8 9. Kh6 +-

White to play

1. b4! creates a passed pawn (1. Ke2? Ke7 2. Kf3 Ke6 3. Kxf4 Kd5=) In the pawn race that follows, both sides queen but White can win by force. However, this is not something that I would care to have to play over the board! 1... Ke7! (1... axb4 2.
a5 +-) 2. b5! Kd6 and the main line, stripped of Rosen's analysis, runs 3. Ke2 Ke6 4. Kf3 Ke5 5. Kg4 Ke4 6. b6 f3 7. Kg3 Ke3 8. b7 f2 9. b8=Q f1=Q 10. Qe5+ Kd2 11. Qxa5+ Kd1 12. Qd5+ Kc1 13. Qc5+ Kd1 14. Qd4+ Kc2 15. Qf2+ +-

One of the most useful chapters for me dealt with separated, isolated, passed pawns. Rosen did an excellent job presenting rules for determining whether such positions are a draw or not.

Black to play

Rosen's rules indicate that here the pawns cannot defend themselves on their own: 1... Kh5! threatens to capture on h4 and still remain in the square of the e4 pawn 2. e5 now 2...Kxh4? would be a mistake, but 2...Kg6! switches the attack to the lead pawn: 3. Kg2 Kf5! 4. h5 Kxe5!=

Here again, Rosen takes basic endgame examples (here: separated, isolated passed pawns) and extrapolates to more complicated scenarios:

White to move

The first stage is a Reti-style manoeuvre (pursuing two objectives at once): 1. Kg3! a5 2. Kf4! a4 3. Ke5! White's king couldn't get into the square of the a-pawn, but by threatening to it got close enough to its own passed pawns. 3...Kg7 (3... a3 4. Kf6! a2 5. g7+! Kxh7 6.Kf7! a1=Q 7. g8=Q+! Kh6 8. Qg6#) 4. Kd4! d5 5. Kc3! +- White can now capture both pawns in the same manner as the previous example.

I hope these examples have whetted your appetite for more. This book is chock-full of positions whose solutions made me think, "whoa, cool!"

In a nutshell, this book impresses me because:

  • it relates basic endgame positions and concepts to more complicated and realistic scenarios
  • it provides rules and guidelines, yet ultimately reinforces the need to think concretely and analyze accurately
  • the careful selection of examples, combined with the chapter "Advice for Teachers", makes this a very good resource for trainers
I highly recommended this book to students that have already studied elementary endgames, and to trainers looking for lecture material.

Tuesday, July 8, 2008

Reti Opening: Rare Bird?

I'm making a push to finish my New York 1924 tournament book annotated by Alekhine. This tournament was a showcase for the Reti opening (18 games, 6 of them classed as the "Reti Reversed").

I updated my file of personal games, and it's approaching 6000 games (mostly ICC). I did a quick search, and only 6 games had Nf3, c4 and g3 as the first 3 moves for white. I did a search of my 3.4 million game database and it only came up about 12000 maybe 3x as often as it does for me, but still pretty rare.

I'm curious if other people find this to be a highly rare opening, either online or over the board. I have no interest in playing the opening as white, and don't think I'd recommend it to a beginner, but it seems like it would be a good surprise weapon based on my statistics.

Saturday, July 5, 2008

Name That Gambit

Here is a quest for my dear readers: does the following gambit have a name? Is there any theory associated with this, or a clear refutation?

1.d4 Nf6 2.Nf3 e6 3.e4:

There were very few examples of this in my large database. I couldn't find a clear path forwards for Black. Surprisingly, Fritz evaluated 3...Nxe4, which I think is the most testing response, to be nearly equal. Even after following plausible moves for both sides down to move 12 or so, Fritz didn't seem to notice that Black was up a pawn! For example, I played this series of moves that felt natural to me and that looked fine to Fritz at first glance: 3... Nxe4 4. Bd3 d5 5. O-O Be7 6. Re1 Nd6 7. Bf4 O-O 8. Nbd2 b6 9. Qe2 c5 10. c3 Nc6 11. Ne5 Bb7:

Fritz evaluates this as dead even, despite the pawn minus. I've only done some preliminary, superficial analysis, and I don't suggest the above variation is some sort of main line. My point is that there seems to be more going on here than meets the eye. It would be interesting to grab another chess engine and see its evaluations, or to play engines against each other and see the results.

Other moves such as 3...c5 and 3...d5 may transpose into other openings such as the Sicilian or the French, but that can't be the most serious test (otherwise this opening would be seen more often).

My conclusion right now is that Black takes the pawn and converts the material advantage with good technique. However, I would say that any unusual pawn gambit that Fritz can't easily refute is probably worthy of home analysis and potential use as a secret weapon.