Friday, February 25, 2011

Once was enough II : Portuguese Gambit

1.e4 d5 2.exd5 Nf6 3.d4 Bg4

Although this series is not going to follow a chronological order, this, the second part, does in fact follow on directly from the first. Having lost the first leg of a two-game match with the Latvian Gambit, I surprisingly won the second and thereby found myself in a rapidplay play-off, again over two legs.

Although the two normal-rate games had taken place six weeks apart, the first of the rapidplay games was arranged for just three days after my second-leg win. The only other chess I played during those six weeks had been a quickplay tournament: I met 1.e4 four times, each time replying with 1...e5 and scoring three losses and a draw. This may have been influential in persuading me to play something different against Adam (though, as we shall see, it may not) and I chose to play the recently-trendy Portuguese Gambit.

WHERE I FOUND IT: I believe I'd seen an article or two about it, which had persuaded me to get hold of a then-new book by Selby Anderson, which made use of a lot of computer analysis (this was not standard practice back in those days) to make a case for it.

In the Portuguese, reckons the back cover,
Black offers a pawn for rapid development and the chance to exploit weak squares in the enemy camp.
Go on...
The resulting play is sharp and trappy, with plenty of "coffeehouse" flavour for those wishing to mix it up with Black.
I don't suppose those are a couple of John Nunn's "secret code words"? The book thinks not:
No mere surprise weapon, however, the Portuguese has been tested in the fire of world class tournament chess. Proof of the line's vitality is it's (sic) adoption by aggressive Grandmasters such as Spraggett, Hodgson, Shirov and Hebden.
I think, now, I might query the term "proof". I might also be inclined to interpret the claim
the high percentage of original analysis found here stands in stark contrast to the typical opening monograph.
as meaning "has lots of untested analysis and very few actual games". But we are all wiser with age, or think we are.

WHY I PLAYED IT: I could claim that I prepared it specially as a one-off opening against a strong player, though given that I'd just suffered a heavy defeat against the same player using much that strategy, it would still have been a strange way to proceed. But in fact, looking back, I'd actually employed the not-dissimilar Icelandic Gambit, 1.e4 d5 2.exd5 Nf6 3.c4 e6!?

twice in the previous fifteen months. As I've never played - nor, as far as I recall, intended to play - 3.d4 Nxd5, I'm pretty sure that against 3.d4 I must have been prepared, for quite some time, to punt a Portuguese. Although 3..Qxd5 is also possible: but if you're going to do that, why not take with the queen on move two?

I did, in fact, play 2...Qxd5 (and 3.Nc3 Qa5) for a while, having bought and liked John Emms' first book on the subject, abysmal production aside. I never entirely believed it though, especially as Anand, who played it against Kasparov to prompt my interest in the first place, absolutely pasted Lautier in a celebrated game just when I was taking it up.

I wrote in the last column:
So, with the logic that was informing my choices at that time, I replaced an intended line that was little-known and a little dubious with one that was even less well-known and even more questionable.
This was exactly the same syndrome. Of course it's easy at a distance to see that this is silly, but harder to see how to avoid it: moreover, if the opening looked potentially busted, one might as well play lesser-known and more complicated lines than go down the main line and hope they don't know Anand-Lautier.

Besides, as I was also still playing 1...e5, I believe the idea was to have more than one line available to me, so that I could vary, depending on the opponent. This is not at all a silly idea, but there are some large questions regarding implementation.

  • Do we really have multiple lines available to us, or are we just hiding the fact that we don't, as yet, have even one we can rely on? (I didn't.)
  • Is there not a tendency for this to get out of hand, so that you end up having several different lines for several different situations, and knowing none of them? (There is: a few years later I found myself trying to keep up with the French, the Sicilian and the open games.)
  • Is there a difference between having a combative line available, and reading up on a wild line requiring you to learn lots of tactical lines that you'll never in fact learn? (There is: and you're all the less likely to learn them if you only play it once.)

The books always talk about how your opponent is unlikely to know the theory: they talk less about how you probably won't, either.

WHY I STOPPED PLAYING IT: Not, especially, because of the result of the game: in fact Black, despite playing, inadvertently, down lines (6...a6 and then 8...Nb6?!) that Anderson dislikes, got splendid compensation for his pawns. (11.Qc4 would have been stronger, as would 12.Ne2, but that's the gamble - sacrifice for some compensation and see if their mistakes can turn it into lots.) Had I not blundered a knight away at move 19, I might easily have gone on to win. But I lost, and the second game was drawn.

I'm not even sure I gave it up straight away: the next time I met 1.e4 I played 1...d5 2.exd5 Nf6, but the opponent played 3.Nc3. After that, I never got as close to playing it again. At the end of the day you can't, really, play dicky lines which require you to learn lots of theory, in pursuit of an opening which you know is probably unsound, and which you will, anyway, never play enough to fix the theory in your memory.

Not unless you really, really like it. I didn't like it nearly enough: I was playing it because I was desperate. And if you're desperate, it's perhaps best to pick a proper opening, and see what happens.

CHANCES OF MY PLAYING IT AGAIN: None. I still like 1...d5: I can imagine myself being tempted to try it again, in its 2...Qxd5 and 3...Qa5 form, though it's probably a temptation better resisted. But the Portuguese Gambit is a one-off opening. And I've already tried it once.

[Once was enough I]

[Once was enough II]


Jonathan B said...

Anderson's book belongs here methinks.

Morgan Daniels said...

I dunno... I've got quite a red/green fetish

Chris Morgan said...

I play the Portugese Gambit, albeit at a lower echelon of the chess circuit than EJH. I often find White doesn't play 4. f3, which is possibly the best move, as it looks a bit awkward. They'll play a simple move like 4. Nf3 or 4. Be2. I think this is a common attitude when facing a gambit that you don't know about in fear of falling into a trap.

Campion said...

My limited experience is similar to that which Chris describes - I spent a few years in the lower echelons of a regional chess league racking up a useful points tally with e4-f4-Nc3-Bc4-and-charge-at-f7 as White, and the Portuguese as Black in reply to 1. e4.

Quite a lot of the time people played 4. Nf3 as a way to avoid spending too much time trying to remember the optimal lines, and play transposed into positions reached after the 2.. Qxd5 lines (with queen on a5, knight on f6, bishop on g4, etc)

Martin Cowley said...

I used to play the Portuguese but gave it up because if White plays one of the quieter lines it is difficult to do more than equalise. I quite liked playing against 4. f3 which is White's best line but is also double-edged. The Anderson book is not bad but there is a lot more up to date theory on it now.