Dear Mr. Silman,

After 1.e4 d5 2.exd5 Nf6 3.Nf3 theory books frown on 3…Bg4. I never exactly understood why. Of course, I understand that White doesn’t have to oblige by transposing into my beloved Portuguese after 4.d4, but I’ve often played 3…Bg4 anyway, when 4.Be2 Qxd5 5.d4 Nc6 6.0–0 0–0–0 7.Nc3 Qh5 8.h3 Bxh3!? has given me 9 wins out of 10 games against players of similar strength (around 1800).

Wondering if this sac could be correct, I let Fritz 7 play a series of 20 min /20 min, engine vs. engine games. I was surprised to see that Black won 7-1! The advance of Black’s K-side pawns – after somewhat lengthy and obscure preparations – proved deadly.

So, my request is:

a) Could someone explain SIMPLY why the seemingly natural move 3…Bg4 doesn’t work after 3.Nf3.

b) If so, Black has to enter a transpositional labyrinth after 3…Nxd5 (I don’t like 3…Qxd5 too much, but it’s just a matter of taste). Is there any good advice about what plan to select? Has an early Nf3 reduced or modified White’s options somehow?

c) What about this sacrifice? Of course, White can avoid it by playing h3 earlier, but then the Bishop just retreats.

Sincerely,

Jean

DOES 3.Nf3 RULE OUT THE PORTUGUESE?

SILMAN comments:

Before going into the position after 3.Nf3, let’s take a quick look at a pure Portuguese and see why so many players enjoy using this attacking system: 1.e4 d5 2.exd5 Nf6 3.d4 Bg4 4.f3 Bf5 5.c4 e6 6.dxe6 Nc6! 7.exf7+? (Very tempting and very bad!) 7…Kxf7 8.Be3 Bb4+ 9.Nc3 Re8 10.Kf2 Rxe3! 11.Kxe3 Nxd4! 12.Qxd4 Qe7+.

A NIGHTMARE FOR WHITE!

Black wins the White Queen and the game via 13.Kd2 Rd8 or 13.Kf2 Bc5, while 13.Kf4? Nh5+ leads to mate. As you can see, this is a FUN opening to play if you enjoy active pieces and crazed tactics!

The annoying 3.Nf3 avoids 3.d4 Bg4 (of course, Black is likely itching to see 3.c4 e6 4.dxe6 Bxe6, which is known as the Icelandic Gambit and has given the second player many attractive victories), which constitutes the starting position of the Portuguese Attack. The real theoretical question is “Can Black achieve a playable game after 3.Nf3 without transposing directly (via 3…Nxd5) into a Scandinavian?”

First, I should point out the fact that Jean sent a bogus (as in illegal) set of moves, so I’ve broken his starting piece sacrifice position down into two move orders:

1.e4 d5 2.exd5 Nf6 3.Nf3 Bg4 4.Be2 Qxd5 5.d4 Nc6 6.h3 Qh5 7.Nc3 0-0-0 8.0-0 Bxh3 (John Watson will use this order). I should point out that in this order, both 6…Bxf3!? (with the follow up 7.Bxf3 Qd7 8.c3 e5) and 6…Bh5!? are far superior to 6…Qh5.

1.e4 d5 2.exd5 Nf6 3.Nf3 Bg4 4.Be2 Qxd5 5.d4 Nc6 6.0–0 0–0–0 7.Nc3 Qh5 8.h3 Bxh3 (the order I gave in Jean’s letter). In this particular move order, Black can play 7…Qf5 (instead of the indicated 7…Qh5) with a satisfactory game. In both move orders, Black should avoid the horrible 8…Bxh3 and instead play 8…Nxd4! with good chances.

Now let’s have JOHN WATSON answer Jean’s query about the piece blunder … hmmm … I mean sacrifice on h3:

1.e4 d5 2.exd5 Nf6 3.Nf3 Bg4 4.Be2 Qxd5 5.d4 Nc6 6.h3 Qh5 7.Nc3

This leads to the reader’s sacrificial line. But I don’t see any way to equality after the alternative 7.0–0!, when the early ...Bg4 hurts Black: 7...Bxf3 (7...Bxh3? 8.Ne5 Qf5 9.Nxc6 bxc6 10.gxh3 Qxh3 11.Bf3 0–0–0 12.Bg2) 8.Bxf3 Qf5!? (8...Qh4 9.Bxc6+ bxc6 10.Qf3 0–0–0 11.Qxc6) 9.Bxc6+ bxc6 10.c4 e6? (10...0–0–0 11.Qa4 Kb7 12.Be3; I think best is 10...Qd7!, but White has various moves that keep a positional advantage, for example, 11.Nc3 e6 12.Re1 Be7 13.Bf4) 11.Qa4 Kd7!? 12.Nd2! Bd6 13.Nf3, threatening c5. This is awful for Black.

7...0–0–0 8.0–0 Bxh3

“MANLY” BUT BAD

9.Ne5

Or 9.gxh3 Qxh3 10.Ng5 Qh4 11.Kg2 (safest? There are alternatives.) 11...Rxd4 12.Rh1! Rxd1 13.Rxh4 and White is winning.

9...Qf5

No better is 9...Bg4 10.Bxg4+ Nxg4 11.Qxg4+ Qxg4 12.Nxg4 Nxd4 (12...Rxd4 13.Nh2 e6 14.Nf3) 13.Ne5.

10.Bd3 Qe6 11.Nxc6 bxc6 12.gxh3 Qxh3 13.Re1 e6 14.Re3 Qh4 15.Rg3 (or 15.Qf3). In my view, these ALL win for White, but you may want to check and improve.

Okay, Silman’s back and will now answer the key question: Can White successfully avoid the dreaded Portuguese Attack with 3.Nf3?

If Black can show that 3…Bg4 forces 4.Be2 then the answer is no, since the game will transpose back into main Portuguese lines. However, 4.Be2 isn’t correct. The only way to punish 3.Nf3 Bg4 is, according to theory, by trying 4.Bb5+.

Emms, in his book, THE SCANDINAVIAN, offers the following analysis: 4…Nbd7 5.c4 (5.h3 Bxf3 6.Qxf3 a6 7.Ba4 b5 8.Bb3 Nb6 9.Nc3 b4 10.Ne4 is strong since 10…Nfxd5 fails to 11.Ng5 f6 12.Ne6 Qd6 13.Bxd5 Nxd5 14.Qxd5 Qxd5 15.Nxc7+, 1-0.) 5…a6 6.Ba4 (Emms also claims a White plus after 6.Bxd7+ Qxd7 7.d4) with a clear plus for White.

Ah, if only life were so simple. After 3.Nf3 Bg4 4.Bb5+ Nbd7 I have to take exception with his fifth move lines:

5.h3!? (Perhaps this is White’s best?) 5…Bxf3 6.Qxf3 a6 7.Ba4 b5 8.Bb3 and now, instead of Emms’ 8…Nb6, I thought Black could improve with 8…Nc5.

I showed this position to Grandmaster Murray Chandler, a part of the powerful writing/editing team at Gambit Publications (www.gambitbooks.com). He replied: “In general, I think that Black has more resources than is generally appreciated in the Portuguese. However, in this line, after 8…Nc5, don’t you think that 9.c4 favors White?”

IS WHITE JUST BETTER?

Okay, I think the grandmaster’s right! Though 8…Nc5 is superior to 8…Nb6, I can’t find a “happy” way for Black to continue after 9.c4. If any readers want to try their hand at this position (many of the lines lead to very strange situations, so who knows what’s lurking under the muck), feel free to send me the results for publication!

For those that want to believe in Black’s Portuguese chances (against all common sense and logic), let me offer another suggestion off the top of my head: 5.h3 a6!? (Note that 5…Bh5 6.c4 a6 7.Bxd7+ Qxd7 8.d4 is just bad for Black) 6.Ba4 (Also critical is 6.hxg4 axb5 7.Qe2. Can anyone make Black’s position work here?) 6…Bh5 7.c4 (Now 7.g4 Bg6 is no longer dangerous since 8.c4 b5! gives Black excellent play.) 7…b5! 8.cxb5 Nxd5 9.d4 (Perhaps 9.Nc3 is stronger?) 9…e6 10.bxa6 Rxa6 11.g4 Bg6 12.Ne5 c6! 13.Bxc6 Rxc6 14.Nxc6 Qa8 and White is toast. Of course, the amount of improvements for both sides along the way is likely staggering.

So 5.h3 is quite testing, but White’s “main” move is supposed to be 5.c4. Now, instead of the oft repeated 5…a6, what’s wrong with 5…e6!? when life suddenly becomes quite interesting: 6.dxe6 (6.Qe2 Be7 is nothing, while 6.Nc3 exd5 7.cxd5 Qe7+! 8.Kf1 0-0-0 is very attractive for Black) 6…fxe6.

DOES WHITE HAVE ANYTHING AT ALL HERE?

This position appears to offer Black some very interesting dynamic possibilities: 7.d4 (7.Bxd7+ Qxd7 8.0-0 0-0-0 pleases the second player, and 7.0-0 c6 8.Ba4 Ne5 is very annoying) 7…c6 8.Ba4 e5 makes for a sharp game in which Black should be doing well.

So, it seems that the theoretically recommended 5.c4 is diffused, but 5.h3 is serious business. The ball is in Black’s court here!

I should also point out that John Watson (who, like myself, admits to not being an expert on the Scandinavian systems) was wondering what was wrong with 1.e4 d5 2.exd5 Nf6 3.Nf3 Bg4 4.d4 Qxd5 (4…Nxd5 takes us into lines that, as Watson later shows, are known to be pleasant for White) 5.Nc3!?.

YET ANOTHER WAY FOR WHITE TO PLAY

He says: “I don’t think that 5...Bxf3 will work; even if you somehow trapped the Knight after 6.Nxd5 Bxd1 7.Nxc7+ etc. (which I doubt), there’s also the ending after 7.Nxf6+ and 8.Kxd1 to deal with. Assuming that’s true, where does the Queen go after 5.Nc3? It’s a bit early to make a decision. 5...Qa5 transposes directly into old main lines that are thought to be better for White (usually Black’s Bishop belongs on f5), 5...Qh5 is a joke, so that leaves 5...Qd6, as in Melts’ book. This is all certainly worth thinking about. Maybe 5.Nc3 is a problem.

FINALLY, for those that were wondering why Black isn’t simply recapturing on d5 with his Knight, here’s a small survey of that line by John Watson.

1.e4 d5 2.exd5 Nf6 3.d4 Nxd5 4.c4 Nb6 5.Nf3 Bg4

This just doesn’t seem accurate to me, but the resulting position is forced when you use the reader’s early ...Bg4 move order (5…g6 being the move of choice among Scandinavian afficianados). This explains why Hansen finds White’s position comfortable.

NOT AN EASY POSITION FOR BLACK

6.Be2

Unfortunately, there’s another big move order problem: 6.c5! is also given as a good line by all the theory that I’m looking at (it seems that everything favors White after 5...Bg4). I won’t go into the theory here. You can find it in any book on the Scandinavian.

6...Nc6

6...e6 may be safer? Probably transposes to the next note.

7.d5

7.Nc3 also has to be considered, usually transposing to a known line after 7...e6 (7...e5 8.0–0 exd4 9.Nxd4 Qxd4 10.Bxg4 may be playable for Black [it’s probably a little better for White], but he has to watch out for lines like 10...Qxc4?! 11.Re1+ Be7 12.Bg5 Rd8 13.Qf3 f6 14.Bh5+ Kf8 15.Rxe7! Nxe7 16.Bxf6 etc.) 8.0–0 Be7 9.d5 exd5 10.cxd5 Nb4 11.Qd4 Bxf3 12.Bxf3 0–0 13.Qd1 – NCO.

7...Bxf3 8.gxf3!

8.Bxf3 Ne5 is probably okay for Black and gives active play, for example, 9.Be2 e6 10.f4 Bb4+ 11.Nc3 Ng6 12.Qb3 Qe7 13.dxe6 0–0!.

8...Ne5 9.f4 Ned7 10.Nc3 Nf6 11.Be3

I think that a real problem is 11.Qb3!?, which looks dangerous. For example, 11...g6 12.Be3 Bg7 13.0–0–0 and White is ready for c5.

11...g6 12.Qc2

Again, 12.Qb3!.

12...c6 13.dxc6 bxc6 14.Rd1

Better may be 14.Bf3 Qc7 15.0–0–0, for example, 15...e6 16.Ne4 Nxe4 17.Qxe4 Rc8 18.Bd4.

14...Qc7, Kirilov-Gipslis, Katowice op 1992. Black stands reasonably well, but White’s options along the way render this line marginal.

CONCLUSION: Black has an awful lot of work to do to revive this ...Nxd5 line after the early (premature?) ...Bg4.

Okay, that’s it. Is the Portuguese still alive? I KNOW there are lots of fans of the Portuguese Attack out there, so this is your chance to be heard.