The appearance of Graham Burgess' excellent book The Taimanov Sicilian inspired me to reopen my ongoing discussion about the modern opening book. In this age of computer databases, we must continually reassess what we expect from an opening book, or at least one that deals with the whole of a mainstream opening. Ideally, such a book might contain: (a) thorough research into databases, books, and other sources, including the use of the increasingly many games annotated by strong players to be found in magazines and databases; (b) a presentation of all standard and plausible variations with assessments, and the author's considered guidance as to which lines are best, which lines have been discarded, and the like; (c) a knowledgeable presentation of strategic themes, typical tactics, and standard positions, aimed at both newcomers to the opening and those who wish to increase their understanding; (d) as much original analysis as possible, hopefully with genuinely original ideas by the author; (e) error-checking of one's own and others' analysis by, but not solely relying upon, computer software; (f) clear and comprehensive indication of transpositions combined with relevant comparisons between similar positions; and (g) entertaining and unstereotyped writing. (With, of course, a pleasing layout and low price).

The problem is that all of this cannot be fit into a reasonably sized volume of, say, 150-200 pages, at least not for any mainstream opening with a lot of history and theory behind it. Thus we hear complaints ranging from a book's lack of explanation for the moves made, incomplete coverage of the opening itself, the shortage of original (non-computer generated) analysis and opinions, too elementary or too advanced a presentation, analytical oversights, and poor writing. An author must make tradeoffs, as is evident in recent efforts of quality. Matthew Sadler deservedly won the BCF Book of the Year award with his Queen's Gambit Declined, which features a well-written and enormously appealing question-and-answer method of instruction. The more I thought about it, the more I appreciated this choice, because it obviously excited and inspired a great number of readers craving clear strategic explanations and a personal touch. Nevertheless, other approaches have advantages as well. Sadler's book isn't what I myself find most useful, because important variations aren't covered, the analysis is necessarily light, and the research limited. It's a matter of taste and tradeoffs. Two of my favorite works from last year, for example, were Steffen Pedersen's Semi-Slav books and John Emms' Play the Open Games as Black, but it must be admitted that they are not as generous or friendly with explanations and are written in a technical style not appealing to everyone. What readers want will certainly vary, but at least we can try to assess how well a book fulfills its aims.

We come to Burgess' The Taimanov Sicilian. This is a 208-page examination of an important and complex opening about which surprisingly little has been written. That in itself counts for something--we've had an awful lot of books and CDs on the Sveshnikov (5 or 6?), Najdorf, and Dragon Sicilians in recent years, for example, and yet only Plaskett's 1997 book on the Taimanov, which gives very limited coverage in comparison with this one.

The Taimanov Sicilian begins 1 e4 c5 2 Nf3 e6 3 d4 cxd4 4 Nxd4 Nc6, normally intending some combination of ...Qc7 and ...a6 to follow. Of course, 5 Nb5 d6 and 6 c4 or 6 Bf4 is also thoroughly examined, and interestingly, Burgess devotes a 9-page chapter to lines in which Black aims for a Scheveningen by means of 5 Nc3 d6. After 5 Nc3, variations with ...a6, ...Qc7, and ...Nf6 are dealt with, as well as those with ...a6 and ...Nge7 (which is called the "Pure Taimanov"). Burgess' book positions itself squarely in the technical camp, and is very strong in areas "a," "b," "e" and "f" above. Let's look at those categories first. Regarding "a," it would be hard indeed to have chosen among more games and analysis, including that from older books and sources, than Burgess has done. His reputation as a researcher precedes him. As for "b," Burgess' grasp on which lines are good or bad and why may be the most outstanding feature of the book. In the midst of very dense and confusing theory, full of transpositions, he not only identifies the theoretically preferred lines and explains why others have fallen into disfavour, but he also directs our attention to numerous underrated ones, reviving them with suggested improvements when necessary. Lines are all error-checked ("e") by computer, but Burgess has also used his judgment to investigate further what the analysis engine considers an "error." Burgess' treatment of transpositions ("f") deserves some discussion. There are an unwieldy number of these in the Taimanov Sicilian, all or almost all of which he seems to find. The method for presenting these is his own invention: the move order to which the material transposes is expressly given in italics, with all the moves necessary to bring about the transposition listed (this sounds obscure, but one catches on immediately). I think this is an improvement (at least for this opening) over the old "12 Nd5 Be6 13 Qd2 transposes to Chapter 7, 'C,' note 'b' to 10 0-0." Upon occasion, Burgess also lists the assessment that is ultimately attached to the line referred to; this is very useful if one doesn't want to waste time jumping back and forth too see if a line is worth transposing to. I wish that he'd used this technique (the assessment) more often, but other authors don't use it at all, so at least we're heading in the right direction. For all that, the ubiquitous transpositions are undeniably distracting. There may be no perfect solution to this problem, but one would certainly rather have every transposition noted than not.

Now let's talk a bit about our other categories. Assessing "c" is rather easy. There's no way around it: Burgess has relatively little strategic explanation (and only a few of those popular diagrams with arrows to show where the pieces might go). He provides some discussion of standard ideas in the chapter about 1 e4 c5 2 Nf3 e6 3 d4 cxd4 4 Nxd4 Nc6 5 Nb5 d6 6 c4, but other chapter introductions are mostly devoted to pointing out which variations he considers best and/or most exciting, not to general strategy. This is a practical approach, allowing readers to begin choosing a repertoire before going on to the relevant lines. Some readers will understandably prefer a more abstract (and verbal) approach to learning. On the other hand, the reader may remember my philosophy that general ideas are best learned through study of the examples in the main line and notes (and especially by comparing very similar lines leading to different results). As I see it, forming one's own generalities through study is just about always superior to having them spoon fed. Thus I am comfortable with Burgess' approach, which incidentally has been the preferred one for most chess books up until the last decade or so. I also think that to the extent that strategic explanation helps to guide the reader's thoughts, it does so most effectively when attached to the lines under discussion, i.e., in the middle of games and analysis of a chapter and not prefatory to it. Probably the book could do with more of that kind of assistance, but that would bring up another tradeoff with completeness and analytical depth. In my experience, a good opening book with a strong analytical bent maintains its usefulness for more years than one with more words and fewer specifics.

Criterion "d" is a tricky one, since originality can manifest itself in different ways. A negative review of this book by Jacob Aagaard, a competent chess author himself, emphasizes his disappointment that Burgess has not been more personally engaged with his material, accepting superficial assessments of certain positions, and that the book is in his view a kind of "mass production." I really disagree with this characterization. In fact, Burgess generates a remarkable amount of original analysis and suggestions for further research, precisely what would be missing in a mechanical approach. Moreover, accepting another annotator's wrong assessment about a position is simply inevitable in such an ambitious project. The question is how frequently this occurs, and how often Burgess challenges or refutes such assessments. I think that any fair reading reveals that he does the latter much more than the typical chess author, and that he regularly extends others' analysis to determine what's really going on. Many of his suggestions are necessarily abbreviated when some further analysis would be helpful, yet they point in a new direction that might otherwise not have been noticed. It's also true that an author has to let a good many games and positions pass by without detailed criticism. Otherwise, given the richness of chess, one would be heavily annotating every example and have no room at all to achieve the kind of complete exposition Burgess seeks.

I honestly believe that these tradeoffs apply to all opening books, including Aagaard's own. His very interesting book on the Panov-Botvinnik Attack (Caro-Kann), for example, has a wonderful chapter on a famous ending that arises from one of the main lines and he does an excellent job of explaining general ideas. But he sacrifices depth throughout; for example, while trying to establish a repertoire for White after 1 e4 c6 2 d4 d5 3 exd5 cxd5 4 c4 Nf6 5 Nc3 e6 6 Nf3 Be7 7 cxd5, he doesn't even consider 7...exd5 (226 games in Megabase 2000), and after 6...Bb4 7 cxd5, he gives only a very superficial look at 7...exd5. Preparing Black to play 5...Nc6, he neglects some simple White options after 6 Bg5 and superficially analyses others. It is, in short, a book with very good explanation, but incomplete coverage, especially for repertoire book. That certainly doesn't make it a bad book--I've used it myself to help teach a student the Panov Attack--but it's an entirely different style of book and whether it's appropriate for you depends upon what your tastes and goals are.

Finally, I think that criterion "g" (entertaining and unstereotyped writing) is very difficult to achieve in so comprehensive a work (how many ways are there to say "Black is somewhat better," "with compensation" or "with the idea of ...Bb4" ?). Given space constraints and the book"s philosophy, I think that Burgess does a fair job of mixing in pertinent comments. Of course, Sadler gets an "A+" in this category, but his book is structured in an entirely different manner.

I used to play the Taimanov (and the related Kan System), and I was impressed with the treatment of traditional lines that I used to have some expertise in. For those with some familiarity with the old "main" line, 5 Nc3 Qc7 6 Be3 a6 7 Be2 Nf6 8 0-0 Bb4 9 Na4, Anand surprisingly played 9...Bd6!? in his just-concluded match with Shirov. This has previously been considered bad due to 10 Nb6! (Shirov played 10 g3 instead). Referring to Burgess book, we find that he has some completely original analysis that may show why Anand was willing to allow 10 Nb6, indicating that 10...Bxh2+ 11 Kh1 Rb8 12 f4 is not at all clear after either 12...Bg3 13 e5 Ne4 (14 Bd3 Bf2!) or 12...Qxb6 13 Nf5 Qxb2.

I was also interested in the idea 5 Nc3 Qc7 6 Ndb5!? Qb8 7 Be3, since that was my own attempt to refute the Taimanov when I was quite young. I remember being very embarrassed when, as I was eagerly showing a friend (and a crowd of onlookers) all of my brilliant ideas after 7...a6 8 Bb6 axb5 9 Nxb5, a master passing by asked me (without hesitation) about 9...Ra5, a move I had somehow not even considered! I was unable to demonstrate anything against it and never looked at the line again. I was in the crowd at the 1999 FIDE World Championship when, to my amazement, Ponomariev played this exact line. It turns out that he and others have some new ideas, so I wanted to see what Burgess would say. To his credit, he does indeed suggest 9...Ra5 as "an idea for the future" although it hadn't been played in any of the 6 games with this line so far. He probably should have analyzed the move further, but at least the reader is directed to an important and as yet unplayed move. By the way, it took my computer only a few seconds to find the line 9...Bb4+ 10 c3 Ba5 11 Nc7+ Kf8? (11...Qxc7) 12 Nxa8 Qxa8 13 Qd6+ Nge7 14 Qa3!, winning material. Sometimes you just have to gnash your teeth about these damned things. Anyway, buoyed by the fact that Ponomariev thinks there's something in it, I analyzed 9...Ra5 in some detail and realized that, despite the material advantage that Black emerges with, his position is extremely difficult to defend. The king on e7 (after 10 Nc7+ Ke7, when I like 11 b4) proves to be much more of a problem than it initially appears.

Returning to Burgess' book, he also makes the alert observation that, instead of 7...a6 in the line above, 7...Nf6 8 f4 d6 9 Qf3 Be7 10 Bd3 0-0 11 0-0 a6 12 Nd4 is perfectly playable for Black. The key is that, instead of 12...Nxd4 13 Bxd4 e5 14 Bb6!, which led to a White advantage in two games, Black plays simply 12...Qc7!, transposing to a normal Scheveningen variation with two extra moves for both sides.

I admit to being prejudiced towards the technical and "complete" approach to opening books, and I therefore find it easy to recommend this one. But even a reader who inclines towards a less technical book will still find that The Taimanov Sicilian provides a valuable resource for this contemporary opening.

YOU CAN FIND THIS BOOK AT