In my column this week I take a look at some opening lines, but classic and contemporary, in which one side (always White, in the examples) makes a move that looks like an obvious, beginner-like tempo-waster (generally Bb5, in a position where the bishop is immediately or soon kicked by a Black pawn) turns out to be a clever way of obtaining an otherwise unavailable benefit. Some of you might find this old hat, but hopefully some readers will find this eye-opening, at least on account of the rich diversity of examples. (Feel free to add more in the comments - there were plenty more I omitted due to space limitations.)
Entries in openings (7)
There is no pending draw death taking place before our eyes in Wijk aan Zee. Going into the round almost 50% of the games (24 out of 49) finished with a winner, and in round 8 today only one game in seven finished in a draw - and it took 55 moves. There has been lots of fire and blood on board, which is just what we the fans like to see.
The tournament leader is Magnus Carlsen, who won his fifth game in a row to reach unshared first with five rounds remaining. His victim today was Baadur Jobava, who has been many players' victim in this event, despite winning in the previous round. Jobava trotted out 1.b3, which is one of his signature openings, only to find himself slightly worse in the opening. With resourceful play Jobava managed to equalize and probably would have drawn if the time control had come a move sooner. In the last moves prior to the control Jobava played rather passively, culminating in 40.Qc1. Maybe Jobava could have drawn with 45.Qf2, but it wouldn't have been easy. Instead he swapped down to a queen ending, and that couldn't be saved as White's king was too weak.
Vasil Ivanchuk shared first coming into the round, but lost a very mysterious game to Wesley So. Ivanchuk had White and followed the Viswanathan Anand - Levon Aronian game from round 1 of the 2014 Candidates; a good idea if all you know is the result of that game, but a terrible idea if you know that a humongous opening improvement was found for Aronian that very day. It was published all around the web and in print, and there have even been a couple of games in the database showing the improvement. (Those games featured very decent players, like Jan Gustafsson.) Somehow Ivanchuk missed all the possible sources showing and even detailing the move, and walked right into it. So was ready, played well, and crushed him. Ivanchuk thus fell a full point behind Carlsen, while So moved into (a tie for) second, half a point behind Carlsen. (He also moved up to #6 on the Live Rating List.)
Another player in (the tie for) second is Maxime Vachier-Lagrave, who crushed Anish Giri in a 4.d3 (Anti-) Berlin. Giri's decision to head for a position where MVL would have an isolated d-pawn doesn't seem to have been a good one, as the enemy bishops received too much scope. From there Vachier-Lagrave turned his attention to Black's kingside, and while Giri managed to hold off the attack it came at the price of a lost rook ending.
Ding Liren also won his game and thereby joined the tie for second. His victim was Ivan Saric, whose decision to play 22...Qxc6 was probably based on a miscalculation. My guess is that he missed the nice tactical trick 27.Nxd5, which netted not only an important pawn but the exchange as well.
Radoslaw Wojtaszek had been tied for first going into the previous round, but with a second straight defeat he's almost surely out of the running. He lost with Black in a 6.h3 Najdorf to Teimour Radjabov after sacrificing a pawn but failing to get enough counterplay in return.
Fabiano Caruana started the tournament with two wins but had gone -2 since then. He badly needed a win, and he got one at Loek van Wely's expense. A win over van Wely turned Carlsen's tournament around; who knows, maybe the same will be true for Caruana. Van Wely started coughing up pawns with White in a sort of Hedgehog, and eventually Caruana managed to convert his material advantage into a win.
Finally, Hou Yifan drew with Levon Aronian in an old-fashioned line of the Giuoco Piano. Aronian tried a little too hard to win, and if White had played 42.Rd6+ she might have had good chances for a win. After Hou's 42.Rxd4 her advantage was too small to win, and Aronian held pretty easily after that.
The games, with my comments, are here. Tomorrow is a rest day, and on Tuesday we'll see these pairings for round 9:
- Saric (2.5) - van Wely (2)
- Giri (4) - Ding Liren (5.5)
- So (5.5) - Vachier-Lagrave (5.5)
- Wojtaszek (4) - Ivanchuk (5)
- Carlsen (6) - Radjabov (4.5)
- Aronian (3) - Jobava (1.5)
- Caruana (4.5) - Hou Yifan (2.5)
In the Challengers' group, it was a bloodbath as usual, though there were "only" five decisive games there today as compared to six in the A-group. Haast beat Gunina (in a surprise), Saleh beat Dale, Navara beat Michiels, Wei Yi beat Klein and van Kampen beat Timman. Navara and 15-year-old Wei Yi are running away with the event, sharing first with 6.5/8; Shankland and van Kampen are next with 5 points apiece.
Update: The game score of the Jobava-Carlsen game was corrupted by an arbiter's error at the end; I've updated and uploaded the correct version in the revised link above.
Or Cohen, A Vigorous Chess Opening Repertoire for Black: Tackling 1.e4 with 1…e5 (New in Chess 2013). 319 pp., €23.95/$26.95. Reviewed by Dennis Monokroussos.
The Petroff (or “Russian”) Defense isn’t a big fan favorite among club players, but its poor reputation is somewhat undeserved. It is solid, yes, and at the GM level it leads to a relatively large number of draws. That’s true, and therefore any GMs reading this are welcome to reject the Petroff if they are unhappy with solid openings that have a slightly higher drawing tendency. For the rest of us, it is at least worth considering the Petroff, as “drawish for GMs” need not mean “drawish for us”.
In fact, the Petroff can give rise to rich and complicated positions. There are lines where Black goes all-out for an attack, where the players castle on opposite wings, and where one player or another sacrifices material for an attack or an initiative. The factors that aggressive chess fans like to see are present in many lines of the Petroff, so that’s one point in its favor.
Another point is that it’s manageable. Once a 1…e5 player meets 2.Nf3 with 2…Nc6, he has to worry about the Italian Game*, the Scotch** and the Ruy*** – three big systems. The Petroff cuts all of that out. That’s not to say that it’s anti-theoretical – it isn’t – but the burden is considerably less than it is after 2…Nc6.
Another important question: how does it fare? In the databases, 2…Nc6 has a better score overall, and is at present more commonly seen in GM and super-GM games. But that’s not to say that 2…Nf6 has disappeared – it hasn’t. Vladimir Kramnik and Boris Gelfand haven’t bothered with it lately, but plenty of other 2700+ players have. In the last few months, the Petroff has been used by Alexander Grischuk (2785), Shakhriyar Mamedyarov (2761), Leinier Dominguez (2754), Ruslan Ponomariov (2751), Anish Giri (2737), Wang Hao (2735), Vassily Ivanchuk (2731), and Wang Yue (2723).
Enough about the Petroff; let’s get to the book. The author, a FIDE Master named Or Cohen, offers a full repertoire based on 1.e4 e5, which means he examines not only the Petroff but White’s second-move alternatives, together with the two types of Four Knights (Scotch and Spanish) that can arise after 3.Nc3. Cohen devotes 80 pages to those non-Petroff possibilities, 140 pages or so to 3.Nxe5 lines, and about 75 pages to 3.d4. He generally presents the material in the complete game format (there are 108 in all, 25(!) of which are his), but there are some short, separate analytical sections as well. Further, the games are all dense with analysis, and once the games pass the point where they are of theoretical significance the coverage rightly thins.
In light of the recently-published mega-monster book on the King’s Gambit by John Shaw, I was curious to see how Cohen’s analysis would stack up against it. Following Mihail Marin’s choice in Beating The Open Games a few years back, Cohen recommends the declined line 1.e4 e5 2.f4 Bc5 3.Nf3 d6 4.Nc3 Nf6 5.Bc4 Nc6 6.d3 Bg4 7.Na4 0-0 8.Nxc5 dxc5 9.0-0 Qd6. After 10.Qd2 Bxf3 11.gxf3 exf4 12.Qxf4 Marin gave 12…Ne5, but Cohen suggests the novelty 12…Nd4 instead. Shaw did not consider this move, so kudos to Cohen so far! The next question, of course, is if it’s good. He considers two moves, 13.Rf2, protecting the c-pawn and keeping the tension, and 13.Qxd6, resolving the tension and undoubling Black’s pawns.
After 13.Qxd6 cxd6 14.c3 Ne2+ 15.Kf2 Nxc1 16.Raxc1 b5 17.Bb3 g6 Cohen recommends that Black play (or try to play) …Kg7, …Nd7 and …f5. Even so, he acknowledges that the ending is slightly better for White, and I’m inclined to agree. One interesting plan for White is 18.Rfd1 Kg7 and now the initially counter-intuitive 19.c4! The point is that after 19…b4 White plays 20.d4, and whether Black takes on d4 or maintains the tension in the center, White will double rooks on the d-file and have a small but long-lasting edge. In my opinion White should be happy with this.
The second option, 13.Rf2, leads to sharper play. His analysis continues 13…b5 14.e5 Qd8 15.exf6 (He doesn’t mention 15.Bxb5, but this may offer White an advantage) 15…bxc4 16.dxc4 Re8 17.Be3 Rb8 18.c3 and now the nice tactical trick 18…Rb6! White cannot take on d4 with the pawn because 19…Rxf6 would win the queen, so 19.Bxd4 cxd4 20.Qxd4 Qb8! Black is down two pawns, but threatens two of White’s pawns and enjoys good counterplay. Cohen briefly considers 21.b3 and 21.fxg7, and rightly states that Black’s compensation is sufficient in both cases. I think 21.c5 is better, and may give White some chances for a small edge. Objectively, though, 13.Qxd6 looks like the best bet, when I would definitely prefer to lead the white pieces.
Of course, this is primarily a book on the Petroff, so how does it fare with respect to that opening? Let’s have a quick look at a couple of lines, starting with the trendy 1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.Nc3 Nxc3 6.dxc3 Be7 7.Be3 0-0 8.Qd2 Nc6 9.0-0-0 Ne5 Cohen’s coverage looks good, combining Konstantin Sakaev’s excellent and prescient analysis (from his 2011 work The Petroff: An Expert Repertoire for Black) with recent games that mostly confirm what Sakaev already said – at least in the main lines. In the sidelines Cohen’s coverage is considerably broader and deeper, so while Sakaev’s book was more impressive in charting the course of future GM play, Cohen’s book seems to do a fair job of covering the current lay of the land while covering more of the loose ends amateurs have to worry about in their games.
Now for the traditional main line: 5.d4 d5 6.Bd3 Nc6 7.0-0 Be7 8.c4 Nb4 9.Be2 0-0 10.Nc3 Bf5 11.a3 Nxc3 12.bxc3 Nc6 13.Re1 Re8 14.cxd5 Qxd5 15.Bf4, and now rather than 15…Rac8, which is the main move by an overwhelming margin, chosen repeatedly by super-GMs like Kramnik, Anand, Gelfand and so on down the line, Cohen opts for 15…Bd6. He writes:
15…Rac8 is Black’s main move in the position. There are 199 games on Megabase 2013 in comparison with 16 with 15…Bd6. [DM: Better numbers: 828 for the former and 76 for the latter in ChessBase’s online database.] I cannot really understand the justification. 15…Rac8 is a passive and awkward move, against which White has many options. (228)
It’s true that things like group-think, fashion and inertia based on no longer viable computer analysis can all contribute to even the world’s greatest players making anti-positional moves, but isn’t a little humility in order? Does he really suppose that Kramnik et al are unaware that unnecessary passivity and awkwardness are bad things? In the Introduction he writes this:
Throughout this book I have made extensive use of Houdini 2.0 pro x64. I think it is only fair to say that a 2400 Fide [sic] master is not privileged enough to make his own assessments of the positions of top players, and neither is he in a position to elaborate on their analyses without the usage of this ‘chess monster’.
In fact his FIDE rating is 2349, and never seems to have passed 2360 (http://ratings.fide.com/id.phtml?event=2806010), but this exaggeration aside what he says is sensible and appropriately humble. In fact, in one way it’s perhaps a little too humble. He’s not a super-GM by any means, but he has played so many games with the Petroff over the years and studied it so thoroughly that his understanding of this opening is probably considerably higher than his rating. Even so, I don’t think it’s enough to justify the comment.
However, Cohen does have an explanation. He cites a 2002 Kasparov-Karpov game where Karpov got into all kinds of trouble before eking out a draw. This strikes me as unlikely: both moves were played at the time – and afterwards too. Further, while it’s true that Karpov was fortunate to draw that game, which was the third in that rapid match, he lost the first game with 15…Rac8! Let’s turn to the details and see what we can find.
After 15…Bd6 16.c4 Qe4 17.Be3 Rad8 18.Ra2 (Kasparov’s novelty) 18…Bg6 19.Qc1 Cohen censures Karpov’s 19…Na5, giving it a question mark and stating that this move “is probably the reason why 15…Bd6 is pushed aside in favour of 15…Rac8” (231). He prefers 19…b6(“!”) instead – as have most players to reach this position since that game. (Incidentally, 19…b6 transposes to a 1992 game between Kamsky and Yusupov, so once again the Kasparov-Karpov story comes up wanting as an historical explanation.) Here two moves have been played, 20.h3 and 20.Rd2, and these are the two moves Cohen considers.
Before we turn to them, let me mention two interesting alternatives. First, in Kasparov’s own analysis in Kasparov vs. Karpov: 1988-2009, he offers 20.c5 Bf8 21.Bb5 as a possibility. Computers have come a ways since 2010, but if I were going to play either side of the position I’d investigate the idea pretty seriously.
Another untried move to consider is 20.Bf1, threatening a nasty discovery against the queen and creating the option of doubling rooks on the e-file. Black has an ingenious defense against this: 20…Qf5 21.c5 bxc5 22.dxc5 Bf8 23.Bb5 and now 23…Ne5! This dirty trick seems to equalize: 24.Nd4 Qc8 25.Bxe8 Nd3 26.Qc3 Nxe1 27.Bb5 and now part two of the dirty trick: 27…Nxg2! White must take the knight and allow perpetual check. That’s not so hard to find with one’s computer in the context of a correspondence game, but to find that line over the board would test most players.
Moving on to the moves that have been tried (and covered by Cohen), let’s start with 20.Rd2. After 20…Na5 21.Qc3 has been played in six email games (Cohen seems unaware of the games, or at least doesn’t mention them or that his proposal is a novelty), and here Cohen gives the untested 21…f6, claiming that Black has counterplay. His suggestion looks reasonable, but it isn’t clear what Black is supposed to be doing to prosecute this counterplay. White will play c5, either immediately or after 22.Bd3 and taking on g6, and it seems that White maintains some pressure and a little initiative. Black has no clear way to resolve the position, and for the foreseeable future must react to whatever White does. His coverage of the other known White try, 20.h3, is a little deeper, but not much and doesn’t offer much by way of guidance once the variation ends.
That said, I think his line holds up. I spent an hour or two looking at different possibilities on move 20 trying to bust his line, using different engines and trying some of my own ideas in addition to the computer’s, and I didn’t manage to prove a meaningful advantage. Perhaps a stronger player (or a player with a stronger computer) will do better than I did, but if your repertoire is good up to at least the IM level that will leave most of you in good shape.
Let’s summarize and conclude, first this part of the review and then the review as a whole. I think his broad comments and historical proposals about 15…Bd6 vs. 15…Rac8 are questionable, and I also think that he could have analyzed the key variations in greater depth. On the plus side, it seems to me that 15…Bd6 is viable indeed, and he is to be commended for his analysis and his independence of mind. Further, those who follow his repertoire (at least for this particular variation) may enjoy the benefit of surprise, as 15…Rac8 is far more popular and well-trodden.
Overall, I have mixed feelings about the book. There’s a lot of material that won’t be part of the repertoire, and even though he pares down the analysis of games once they finish the theoretical phase there are still 108 games in the book. Even if each game finishes up in just a quarter of a page it’s still 27 full pages that could have been given to theory or to further explanation. (My guess is that it’s closer to 40 pages in total.) As for the analysis, my feelings are mixed here too. What I checked of his Petroff analysis held up to scrutiny, but his line against the King’s Gambit (above) and the Bishop’s Opening/Italian game (see below) is less convincing. In general, his suggestions – both the good ones and the ones that are less clear – seem to be a bit sketchy. (Again, an argument for fewer games or a different format.) The book could have been better, but if you play the Petroff you should get your paws on a copy, no question. [Note to New In Chess advertising types: please don’t take the part of the last sentence starting with “if” and turn it into a marketing blurb.] If you don’t play the Petroff but are curious about it, it’s again a worthwhile book. For anyone who buys the book, I’d recommend some caution when it comes to the non-Petroff parts of the repertoire. If something catches your fancy then by all means give it a try; but don’t take his Petroff expertise as a reason to switch everything else.
* This (the claim that playing the Petroff allows one to sidestep the Italian Game) is true in principle, but as Cohen recommends meeting the Bishop’s Opening (1.e4 e5 2.Bc4) with 2…Nf6 3.d3 Nc6 (rather than 3…c6, which scores at least as well as 3…Nc6) 4.Nf3 Bc5, someone following his repertoire must know at least one important line of the Italian Game as well. About this, he is surprisingly dismissive. After 5.c3 0-0 6.Nbd2 (no mention of 6.Bb3 with Be3 before Nbd2) 6…a6 7.Bb3 Ba7 8.h3 (no mention of 8.Nf1, used, for example, by Magnus Carlsen in an impressive win over Hikaru Nakamura in the London Chess Classic in December of 2011) 8…0-0 9.Nf1 Be6 10.Ng3 he gives 10…h6, with the idea of …d5, and stops there. (That’s not his main line, but his absolute main line is very bad for White.) And maybe he’s right, but White’s score in the database is excellent here, and at the GM level Black has been getting pounded.
Let’s take another of his lines: 5.0-0 0-0 6.c3 d6 7.Bb3 h6 8.Nbd2 a6 9.h3 Ba7 10.Re1 Nh5 11.Nf1 Qf6 12.Be3 Bxe3 13.Nxe3 Ne7 14.d4 Nf4. Cohen stops here, asserting that Black has counterplay. At first glance that seems a reasonable assessment, as one looks at various sacs on h3 and imagines the black queen sliding to g6. After 15.h4, however, the promised counterplay looks a bit sketchy, my engines (Komodo TCEC and Houdini 4) think White has an edge (or very nearly so), and in the database White’s score is +4-1=3, of which all but one game – won by White – was played by correspondence. This doesn’t at all mean that his analysis is bad, only that it’s a bit sketchy. To some extent it’s unavoidable when trying to cram so much into a single volume, but I do think that the Italian game with d3 + c3 is sufficiently important to deserve more care than it received. John Emms wrote an entire book on this system in 2010 (Beating 1 e4 e5: A Repertoire for White in the Open Games – not given in Cohen’s bibliography), so Cohen’s attempt to dismiss it in less than a single page is rather dubious.
** The “real” Scotch (1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4, and in case of 4…Nf6 5.Nxc6 rather than 5.Nc3) is avoided by Petroff players, but not the lesser, semi-toothless Scotch Four Knights (1.e4 e5 2.Nf3 Nf6 3.Nc3 Nc6 [Black’s last two moves are played in reverse order in the usual scheme of things] 4.d4).
*** Likewise, the Spanish Four Knights can arise via a normal Ruy (as an anti-Berlin line), but only rarely does. Whatever the line’s merits, it shouldn’t really be associated with a normal Ruy.
Inspired by some of the mega-prep at Wijk aan Zee this year, especially the 79-move "speed chess" game Nepomniachtchi-Shirov from round 5, I thought back to another bit of seemingly flawless home cooking by Shirov. In Madrid 1996 he drew a 42 move game with Azmaiparashvili that had been worked out in advance all the way to the end. In Fire on Board Shirov drolly concluded his commentary on the game "Sometimes the Botvinnik variation gets so boring", and with that the line chosen by Azmai was buried.
Very impressive, but before you bury your head in anti-theory or take up Chess960, learn a lesson from Loek van Wely. He took a more careful look at Shirov's idea, and more importantly, thought deeply about what it was intended to achieve. By varying just before Shirov's own improvement, he was able to thwart Shirov's idea, and now the burden of proof is on Black to stay alive in that variation.
There are a lot of lessons to be learned from this game, and the game itself is pretty good too! To see it, and hear some further ruminations on the matter, have a look here. The show is free, as always (free registration required) and will be available on-demand for the next month or so.
Here's the quote:
I also heard a remarkable opinion from another Top 10 player (again I do not have permission to name him) who claimed that my book was a "crime" against my colleagues, since now it is very easy for amateurs to pose problems against grandmasers!
Who wrote this, and more interestingly, who is the Top 10 player? It would be nice to know, so I can ignore or at least quadruple-check anything he has to say about the opening.
Sitting alone, all by itself at the end of this week's issue of TWIC, was a corrected version of a game Epishin (2607) - Schoeneberg (2269) from the Porzellancup rapid about three weeks ago. The game began 1.d4 Nf6 2.Nf3 e6 3.Rg1 (hmm, wonder what he's up to) 3...c5 4.c3 d5 5.g4. While this idea isn't losing for White, Black is probably a little better by this point. Further, White's Rg1 + g4 creates all kinds of possible long-term problems, and when Epishin slipped later his opponent crushed him. Epishin may be a very strong and experienced GM, but when you play junk even the white pieces, 300+ rating points and your opponent's age (Schoeneberg is 63 or 64) may not save you.
There are lots of free downloads at the internet archive, including lots of chess books. Not all of the books there are especially interesting, and some (e.g. all four entries of Soviet Chess by Wade) turn out not to be of chess at all, but it's still worth a browse. (HT: Tim Cianciola.)
One entry I found interesting and then amusing was Frank Marshall's Marshall's Chess Openings. Some things he says there look reasonable, and a lower club player can get some good general ideas from the book. On the other hand, it's pretty funny to read claims like 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 is slightly better for Black, or that Black's best defense to the Ruy is 3...f5.
At a deeper level, I found the book provocative. My first reaction, especially upon seeing such categorical remarks, was to laugh at how primitive opening theory was at the time - even considering that he was writing for the general public at a time when amateur play was far more casual. But then I thought about some of my games against average club players, and realized how thin their opening knowledge really is, too, most of the time; especially when they're not in a pet opening. (One memorable tournament occurred in 2004 when, incredibly, in 6 of my 7 games I had a significant advantage by move 6!) Maybe there's a place for such primitive books in chessplayers' libraries after all. They can outgrow them, and hopefully quickly, but maybe it's a place to start.