A Short Review of Simon Williams' The New Sicilian Dragon
Simon Williams, The New Sicilian Dragon. Everyman Chess, 2009. Reviewed by Dennis Monokroussos.
Maybe it’s something in the water in Great Britain, but those guys publish material on openings faster than politicians make promises and tell lies. Between Everyman, Gambit, ChessPublishing.com, Batsford and Quality Chess, British opening analysis is being delivered to a public with seemingly infinite discretionary income so fast they’re going to have to invent new openings to keep them from getting a day off. (I sometimes wonder whether “Richard Palliser” is an actual person or the name of a secret group, like “Bourbaki” in mathematics. If it is an actual person, I hope they at least give him bathroom breaks. He is so prolific there might be a market for “Richard Palliser facts” over at Everyman, akin to those invented for Chuck Norris. But I digress...)
The book under consideration in this review is Simon Williams’ The New Sicilian Dragon, which claims to be the first book (at least in English) primarily devoted to the “Dragadorf” Sicilian. In case you’re not sure about the variation I’m referring to, it generally arises like this: 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 and now 6…a6, creating a sort of Dragon/Najdorf hybrid. (Thus “Dragadorf”; but because it’s more Dragonlike than Najdorfish, “Dragondorf” seems more appropriate – plus it sounds better.) This variation is not new: Botvinnik played it a couple of times (this I knew) and so did Alekhine (this I didn’t). Despite this pedigree, the variation has only recently captured a broader audience. Many grandmasters have tried it, including such Dragon specialists as Chris Ward, Sergey Kudrin and Mikhail Golubev, and in the upper reaches of the atmosphere Ivan Cheparinov and Alexander Khalifman can be counted among its (occasional) adherents. It isn’t hugely popular yet, but that’s in part due to its lack of publicity and somewhat amorphous nature; this book takes a step towards fixing both impediments.
Simon Williams is a British Grandmaster who started playing the Dragadorf in 2004, so this is not publishing for its own sake; he has put his money where his mouth is, and knows whereof he speaks. His results with it against his peers have been respectable, and when he plays down, he is more than capable of showing the line in its glory. Here’s a game from this past July, played too recently for his book:
T. Heinatz (2294) – S. Williams (2527), Swiss Championship 2009:
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 g6 6. Be3 a6 7. f3 Nbd7 8. Qd2 Bg7 9. O-O-O b5 10. Kb1 Bb7 11. Bh6 Bxh6 12. Qxh6 Rc8 13. Bd3 Rxc3 14. bxc3 Qc7 15. Ne2 Nb6 16. g4 Qc5 17. Ka1 Na4 18. Qd2 Nd7 19. Rb1 Ne5 20. Rhf1 O-O 21. h4 d5 22. h5 dxe4 23. fxe4 Nxg4 24. hxg6 hxg6 25. c4 Qe5+ 26. c3 Rd8 0-1
About this game, it can be said that if Heinatz had and examined Williams’ book, he would not have lost like this! (For starters, Williams argues that Kb1 in this particular variation is often a waste of time.) Interestingly, though, Williams mentions 15…d5!? and 15…Nc5 16.g4 Na4 in his notes to the game J. Ibarra Jerez – M. Vasiliev, Salamanca 2005 (game 20 in the book, in which Black chose 13…e5 instead) but not the move he actually played. This doesn’t strike me as any sort of secretiveness on his part, though, as both 15…Nc5 and 15…Nb6 are nearly equivalent – the main point of both moves is to route the knight to a4.
The book has quite a lot going for it, not least because it currently has the field all to itself. The games seem well-chosen, and Williams does a good job of laying out the main plans and presenting a full repertoire. Let’s start with a look at the contents:
In chapter 1, he covers the position after 6.Be3 Bg7 7.f3 a6 8.Qd2 Nbd7 9.Bc4, calling this the critical test of the Dragadorf at the time of his writing. He notes that Black can choose between four different plans: (1) Active counterplay based on the …Rxc3 sacrifice; (2) striking out with …e5 and …b4; (3) an early …Nc5 and …Nxb3, often combined with 0-0-0; (4) Black plays an early …h6.
In chapters 2 and 3, White varies with 9.0-0-0, and just as in the main line Dragon Black’s counterplay is often based on …d5. In chapter 2 White goes for a quick g4-g5 (with the h-pawn often following along), while in chapter 3 White continues with 10.Bh6 (after 9…b5) before shoving the kingside pawns. Against this latter idea, Williams notes two different approaches: ones with …Rxc3 and those with …b4.
In chapter 4 we see the “positional approach”. Instead of 9.Bc4 or 9.0-0-0, White tries 9.g4 b5 10.g5 Nh5 11.a4. It’s a little unusual at first to think of g4-g5 as “positional”, but White is trying to misplace Black’s pieces rather than blow him off the board with an attack. Williams also examines the early deviations 10.Nc6 and 9.Nd5.
In chapter 5 we find the Accelerated Dragadorf with 6…a6. The pluses are that it saves a tempo if White plays the Qd2 + Bh6 plan, while the earlier …b5 makes it harder for White to play Bc4. On the other hand, pushing the b-pawn before White commits to castling long makes the a4 rejoinder positionally dangerous for Black. Naturally, Williams focuses his attention on this last possibility.
Chapter 6 takes a step back toward “normal” Dragon lines in reply to White approaches like 6.Bc4 or 6.Be3 + 7.Be2. Even here, though, it’s is possible for Black to Dragadorf his opponent, as Short did against Glek (successfully) and Anand (unsuccessfully) back in the mid-to-late 90s.
Finally, chapter 7 leaves the Dragadorf behind and examines other non-6.Be3 tries like 6.g3, 6.f4 and 6.Bg5. Once you’ve reached the Dragon on move 5, Williams has you covered.
I like the structure of the book and much else besides, and those who play this or are interested in playing it should at least consider picking it up. The book includes games through 2008, but as this isn’t yet the world’s most topical line there’s still time to catch up and supplement the material.
So far, so good. I was playing through the games in the e-book, without an engine running, just trying to pick up the general ideas, and then something caught my eye. In game 12, M. Perez Candelario-I. Cheparinov, Dos Hermanas 2005, he presents a variation (starting with 22.Rd7! instead of the game’s 22.Qxg3?) and concludes it with this comment: “White's attacking chances look very good. Indeed, it's quite possible that White is just winning.” The statement provoked my curiosity, so I switched on Rybka, which instantly gives White a nearly four pawn advantage (which doesn’t decrease as the engine continues to run). That left me wondering if his comment was British drollery, or if instead Williams’ use of the computer was at best sporadic. One should not use the computer to analyze when one is training, but definitely should when creating an opening book! Likewise, later in the game, when examining 45…Qd5 (instead of Cheparinov’s 45…bxc4??) he considers 46.Qd2 and 46.Qe3, but for Rybka 3 46.Qb4+ is always its main move. While these points are no theoretical significance whatsoever, I started to wonder if they were indicative of a relative lack of computer use in checking his analysis. That might well be very serious in the context of the sharp lines in the opening.
So, from here on out I decided to let the engine run while I went through the rest of the games, to see how his judgments squared with the silicon monster’s in the opening. In the next game, he offered an improvement for Black which was quite sensible – no problem there – but within that variation clear improvements were available to both players. Later in the game (post-theory), he offers an improvement for Black, but suggests it’s still going to see tough times after White’s rejoinder. Unfortunately, the move he suggests for White is a blunder, allowing Black to draw. (His basic assessment was correct, however, and with other moves White can keep a significant advantage.)
In game 14 (Swinkels-Chatalbashev), however, he explicitly refers to a possible White improvement as “the computer’s suggestion” – a relief! Ironically, there are several moments in the subsequent analysis of the computer’s suggestion where Williams’ analysis differs fairly significantly from Rybka’s. Later in the game, on move 19, he considers one Black option dubious, concluding at the end of the variation that “Black has a problem with his king.” This is indeed the case by the end of the line – White is winning – but only because the last move given for Black is a blunder. After a better move (21…c3 rather than 21…e6, in case anyone is curious), Black is better, perhaps considerably so.
In game 15, the analysis looked more reliable. There were some moments where he slightly disagreed with the computer, but it was clear that he had taken it into account and I tended to agree with his assessments. There were a few moments where I disagreed, but perhaps these were more human disagreements than anything else. Game 16, there was one moment where I think he misevaluated the position, but generally it was clear that he consulted with the oracle.
I went through a few more games, but rather than report all my results, I’ll leave you with a simple caveat emptor. We all err, and the computer’s outputs aren’t infallible either. Even so, there were more gaps than I would have liked to see, but overall I think the book has much more good than bad. Just make sure you check his suggestions with your engines before you trot them out in tournament play!
Maybe it’s something in the water in Great Britain, but those guys publish material on openings faster than politicians make promises and tell lies. Between Everyman, Gambit, ChessPublishing.com, Batsford and Quality Chess, British opening analysis is being delivered to a public with seemingly infinite discretionary income so fast they’re going to have to invent new openings to keep them from getting a day off. (I sometimes wonder whether “Richard Palliser” is an actual person or the name of a secret group, like “Bourbaki” in mathematics. If it is an actual person, I hope they at least give him bathroom breaks. He is so prolific there might be a market for “Richard Palliser facts” over at Everyman, akin to those invented for Chuck Norris. But I digress...)
The book under consideration in this review is Simon Williams’ The New Sicilian Dragon, which claims to be the first book (at least in English) primarily devoted to the “Dragadorf” Sicilian. In case you’re not sure about the variation I’m referring to, it generally arises like this: 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 and now 6…a6, creating a sort of Dragon/Najdorf hybrid. (Thus “Dragadorf”; but because it’s more Dragonlike than Najdorfish, “Dragondorf” seems more appropriate – plus it sounds better.) This variation is not new: Botvinnik played it a couple of times (this I knew) and so did Alekhine (this I didn’t). Despite this pedigree, the variation has only recently captured a broader audience. Many grandmasters have tried it, including such Dragon specialists as Chris Ward, Sergey Kudrin and Mikhail Golubev, and in the upper reaches of the atmosphere Ivan Cheparinov and Alexander Khalifman can be counted among its (occasional) adherents. It isn’t hugely popular yet, but that’s in part due to its lack of publicity and somewhat amorphous nature; this book takes a step towards fixing both impediments.
Simon Williams is a British Grandmaster who started playing the Dragadorf in 2004, so this is not publishing for its own sake; he has put his money where his mouth is, and knows whereof he speaks. His results with it against his peers have been respectable, and when he plays down, he is more than capable of showing the line in its glory. Here’s a game from this past July, played too recently for his book:
T. Heinatz (2294) – S. Williams (2527), Swiss Championship 2009:
1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 g6 6. Be3 a6 7. f3 Nbd7 8. Qd2 Bg7 9. O-O-O b5 10. Kb1 Bb7 11. Bh6 Bxh6 12. Qxh6 Rc8 13. Bd3 Rxc3 14. bxc3 Qc7 15. Ne2 Nb6 16. g4 Qc5 17. Ka1 Na4 18. Qd2 Nd7 19. Rb1 Ne5 20. Rhf1 O-O 21. h4 d5 22. h5 dxe4 23. fxe4 Nxg4 24. hxg6 hxg6 25. c4 Qe5+ 26. c3 Rd8 0-1
About this game, it can be said that if Heinatz had and examined Williams’ book, he would not have lost like this! (For starters, Williams argues that Kb1 in this particular variation is often a waste of time.) Interestingly, though, Williams mentions 15…d5!? and 15…Nc5 16.g4 Na4 in his notes to the game J. Ibarra Jerez – M. Vasiliev, Salamanca 2005 (game 20 in the book, in which Black chose 13…e5 instead) but not the move he actually played. This doesn’t strike me as any sort of secretiveness on his part, though, as both 15…Nc5 and 15…Nb6 are nearly equivalent – the main point of both moves is to route the knight to a4.
The book has quite a lot going for it, not least because it currently has the field all to itself. The games seem well-chosen, and Williams does a good job of laying out the main plans and presenting a full repertoire. Let’s start with a look at the contents:
In chapter 1, he covers the position after 6.Be3 Bg7 7.f3 a6 8.Qd2 Nbd7 9.Bc4, calling this the critical test of the Dragadorf at the time of his writing. He notes that Black can choose between four different plans: (1) Active counterplay based on the …Rxc3 sacrifice; (2) striking out with …e5 and …b4; (3) an early …Nc5 and …Nxb3, often combined with 0-0-0; (4) Black plays an early …h6.
In chapters 2 and 3, White varies with 9.0-0-0, and just as in the main line Dragon Black’s counterplay is often based on …d5. In chapter 2 White goes for a quick g4-g5 (with the h-pawn often following along), while in chapter 3 White continues with 10.Bh6 (after 9…b5) before shoving the kingside pawns. Against this latter idea, Williams notes two different approaches: ones with …Rxc3 and those with …b4.
In chapter 4 we see the “positional approach”. Instead of 9.Bc4 or 9.0-0-0, White tries 9.g4 b5 10.g5 Nh5 11.a4. It’s a little unusual at first to think of g4-g5 as “positional”, but White is trying to misplace Black’s pieces rather than blow him off the board with an attack. Williams also examines the early deviations 10.Nc6 and 9.Nd5.
In chapter 5 we find the Accelerated Dragadorf with 6…a6. The pluses are that it saves a tempo if White plays the Qd2 + Bh6 plan, while the earlier …b5 makes it harder for White to play Bc4. On the other hand, pushing the b-pawn before White commits to castling long makes the a4 rejoinder positionally dangerous for Black. Naturally, Williams focuses his attention on this last possibility.
Chapter 6 takes a step back toward “normal” Dragon lines in reply to White approaches like 6.Bc4 or 6.Be3 + 7.Be2. Even here, though, it’s is possible for Black to Dragadorf his opponent, as Short did against Glek (successfully) and Anand (unsuccessfully) back in the mid-to-late 90s.
Finally, chapter 7 leaves the Dragadorf behind and examines other non-6.Be3 tries like 6.g3, 6.f4 and 6.Bg5. Once you’ve reached the Dragon on move 5, Williams has you covered.
I like the structure of the book and much else besides, and those who play this or are interested in playing it should at least consider picking it up. The book includes games through 2008, but as this isn’t yet the world’s most topical line there’s still time to catch up and supplement the material.
So far, so good. I was playing through the games in the e-book, without an engine running, just trying to pick up the general ideas, and then something caught my eye. In game 12, M. Perez Candelario-I. Cheparinov, Dos Hermanas 2005, he presents a variation (starting with 22.Rd7! instead of the game’s 22.Qxg3?) and concludes it with this comment: “White's attacking chances look very good. Indeed, it's quite possible that White is just winning.” The statement provoked my curiosity, so I switched on Rybka, which instantly gives White a nearly four pawn advantage (which doesn’t decrease as the engine continues to run). That left me wondering if his comment was British drollery, or if instead Williams’ use of the computer was at best sporadic. One should not use the computer to analyze when one is training, but definitely should when creating an opening book! Likewise, later in the game, when examining 45…Qd5 (instead of Cheparinov’s 45…bxc4??) he considers 46.Qd2 and 46.Qe3, but for Rybka 3 46.Qb4+ is always its main move. While these points are no theoretical significance whatsoever, I started to wonder if they were indicative of a relative lack of computer use in checking his analysis. That might well be very serious in the context of the sharp lines in the opening.
So, from here on out I decided to let the engine run while I went through the rest of the games, to see how his judgments squared with the silicon monster’s in the opening. In the next game, he offered an improvement for Black which was quite sensible – no problem there – but within that variation clear improvements were available to both players. Later in the game (post-theory), he offers an improvement for Black, but suggests it’s still going to see tough times after White’s rejoinder. Unfortunately, the move he suggests for White is a blunder, allowing Black to draw. (His basic assessment was correct, however, and with other moves White can keep a significant advantage.)
In game 14 (Swinkels-Chatalbashev), however, he explicitly refers to a possible White improvement as “the computer’s suggestion” – a relief! Ironically, there are several moments in the subsequent analysis of the computer’s suggestion where Williams’ analysis differs fairly significantly from Rybka’s. Later in the game, on move 19, he considers one Black option dubious, concluding at the end of the variation that “Black has a problem with his king.” This is indeed the case by the end of the line – White is winning – but only because the last move given for Black is a blunder. After a better move (21…c3 rather than 21…e6, in case anyone is curious), Black is better, perhaps considerably so.
In game 15, the analysis looked more reliable. There were some moments where he slightly disagreed with the computer, but it was clear that he had taken it into account and I tended to agree with his assessments. There were a few moments where I disagreed, but perhaps these were more human disagreements than anything else. Game 16, there was one moment where I think he misevaluated the position, but generally it was clear that he consulted with the oracle.
I went through a few more games, but rather than report all my results, I’ll leave you with a simple caveat emptor. We all err, and the computer’s outputs aren’t infallible either. Even so, there were more gaps than I would have liked to see, but overall I think the book has much more good than bad. Just make sure you check his suggestions with your engines before you trot them out in tournament play!
Posted by Dennis Monokroussos on
Wednesday August 19, 2009 at 5:20pm
Love the book reviews! I know they're a royal pain, just wanted to let you know how much they are appreciated.
Best,
-Marc-
Do you mean that, after 45...Qd5, 46.Qb4+ is the best move (which maybe challenges author's assessment of the 45...Qd5 line?). Or do you mean that it was necessary to analyse this move in the book because it is the first line according to Rybka3?
As far as I see after 45...Qd5! Black is better whatever White is doing, while 45...bxc4? as Cheparinov played led to a draw by force, so what is a point? (Possibly, 46.Qd2 and 46.Qe3 are analysed because these moves require more precision for Black, or whatever).
Regarding 22.Rd7 I do not see contradiction. No one can be obliged to say that such a position is "completely winning" or anything like that.
In order to suggest that the author might have used engines unsufficiently, something else is needed: flaws in the theoretically significan lines, lots of them.
I appreciate your concern on behalf of Williams and chess authors everywhere. Since I'm on vacation and don't have access to Williams' book at the moment (or the inclination to spend time analyzing), I'll make a few general points.
First, my opinion of the book is basically positive, but if I see something that concerns me, it's appropriate that I mention it. It seems to me a responsibility if I'm telling others that this is a worthwhile way for them to spend their chess money.
Second, the structure of the argument is not "Williams failed to use a computer on some obvious points, therefore you can't trust him." Rather, it's this: "There are consistent indications that Williams failed to use a computer, at least in the sample I examined, and not only in parts of the game that aren't of theoretical significance but in the opening as well. Therefore, the reader should beware and check the details."
This seems reasonable to me. Of course, all authors, even if they use chess engines, are fallible, and it's also a good idea for readers to take responsibility and check anyway. But shouldn't the reader expect that an author would have computer checked his entire work, and certainly the theoretical parts of an opening book?
Just to reiterate, though: I don't claim that Williams doesn't use a computer at all, or that the book is rife with errors, or that players should avoid it.