Is the Portuguese Gambit playable?

I was wondering if the Portuguese Gambit of the Scandinavian Defense was playable.  I know that many strong players don't believe in it, but Ukrainian Grandmaster Sergei Krivoshey plays it sometimes.  Thanks in advance for you responses.

Anyone?

I have seen it played and it seems very playable. Generally you opponent will play 4. Nf3 and after Qxd5 this is, of course, playable. The only way to exploit this move order to prove it "un-playable" would be 4.f3. However, after 4. ...Bf5 Black seems fine and should get enough compensation if he tries to hold on to the pawn. Looks solid to me.

It also depends what you mean by "playable". The move that scores best for white is 4.Bb5+, White's stats here at GM level are much better than in "normal" openings so the opening would probably not hold up as a regular defense in say, a World Championship contest.

At ordinary GM level and below, it's no doubt OK.

At 1600 level, more or less anything is playable, even openings considered theoretically "dubious". 

Thanks guys, NM zkman, is White doing better with 4. f3 or 4. Bb5+?  I've seen many games by the above Grandmaster where the opponent played 4. Bb4+ as well.

You already know the answer. It's been played repeatedly by some very strong players. It has even been played in high level correspondence games, so yes it's playable.

I wouldn't recommend it to most players mostly because there are many openings that give Black better chances to win without having to study a specialized variation. White needs only spend a short amount of time to come up with a playable plan whereas Black will need to spend more energy to make sure they don't slip into a bad position. 

There is an extremely problematic line, which was analysed in some detail in Kaissiber and discussed at chesspub forum. Here is the thread:

http://www.chesspub.com/cgi-bin/yabb2/YaBB.pl?num=1309336195

White may not enter this mess, and seek a simple&slight positional advantage instead.

Playable? Well, yes. Good? Certainly not.

Pfren, I'm familiar with that thread. I agree that White has several different paths to a fully playable game. That thread shows though, that Black is not actually lost.

I wouldn't mess with it as Black, even in correspondence. But you're right. There are some excellent reasons for Black to avoid this.

Thank you all for your responses.  I have decided I'll only use this variation if I'm seeking simply to surprise my opponent, or in a blitz game.

It's my main blitz reply to e4 at the moment.

Bb5+ is the most problematic and I find Be7 is just losing in practice.

Nbd7 however is more playable IMHO.

Here is the main line.

Houdini 3 (22 moves deep) gives it as +0.23 to White after bxa6 Rxa6 Ne2.

The Fritz Engine oddly seems to rate White as much better (0.7+) but I'd rather trust Houdini.

I expect that Fritz is overweighting White's passed a pawn somewhat. Playing through the recommended moves with Fritz is rather odd as the assessment climbs to over +1.00 for white (usually a won game with an engine) but then falls off to half that as you follow the recommended line). Anyway it's clear there is still plenty of play left over the board.

All the engines agree that Black equalises (and in fact may have slight edge) after the natural 9. Nc3.

Black also has the option of having the white bishop on b5 rather than a4 if he prefers - as he can take on b5 on move 8 and then play Nxd5 after white recaptures on b5.

For example here is an interesting continuation following the 8. ...axb5 option for Black. Note that the Nc2 'threat' at move 10 is actually pretty hollow but I can imagine it would suck up a lot of White's time over the board.

What do you do on 5.Be2 if I may ask?

I think I exchange bishops then take the pawn off.

I get a cramped game but no worse than playing the mainline Scandy with no gambits.

A typical engine game I just tried went like this:

With Houdini saying position is = (actually about 2/10ths of pawn up for White).

The main alternative on move 6 is to retake with the N.

That game with engine best moves at blitz speed went:

Again with the engine rating the game equal at the end.

Wait, what do engine blitz games prove exactly?

Well given that none of us are GMs or likely to play better thab Houdini even at 10 secs per move, they tend to demonstrate whether or not a line is 'playable' as asked in the thread title.

You can certainly get more space and a comfortable game as White if you know exactly what you are doing, but the Portuguese has not been busted and in practical terms is perfectly playable. This is especially the case as it's such a rare line.

Perhaps people don't play like engines though?

Certainly you can have lines that need to be played so accurately by one side to equalise that the fact an engine can find its way to equality is irrelevant. That isnt the case in these lines though. The chances of going astray by making natural moves seems more or less the same on both sides. It doesn't feel like Black is walking a tightrope so to speak.

It's a two result position in any event, hardly an accomplishment for Black

Expertise87,

1. The OP wanted to know whether it was playable for Black, not if it was good.

2. Black seems fine in both positions, Houdini rating it as equal seems to vindicate my opinion...