However in the game i managed to hold the advantage and get the pawn pushed to promotion.
My first impressions of Halloween Gambit

Gambits are supposed to be fun and exciting, trading material for time. If GMs avoid them, it isn't always due to soundness. The Halloween Attack may (or may not) be unsound, but at levels below master, black has a hard time proving this, even in correspondence games. So, as a practical gambit, it's fun which is one of the major criteria. Since most gambits offer material for attack potential, one way to sometimes dull that attack is to give back the material. That's the idea behind this supposed refutation to the Halloween Attack:

batgirl, don't dissapoint me, gambits like the King's Gambit or Evans or even Smyth Morra are much closer to being almost sound than this crap. The Halloween is a computer generated gambit.

Well, my purpose is neither to appoint nor disappoint you. But for peace of mind, I had provided the well-known "refutation" in my previous posting The Halloween attack isn't computer generated. It was an often-used opening at the Leipzig Chess Club way back in the 1880s, hence it's other name, the Leipzig Gambit. But the Leipzig players called it the Mueller-Schultze Gambit.
One of the leading experts in the Halloween Gambit (and who plays it in tournaments) is Grigor Minchev:
I've been searching in vain for Tim Krabbé's original article, from many years ago, at Chess Curiosities where he discussed his first, and unpleasant, contact with the Halloween Gambit but I did find some later articles on the opening by Krabbé':
http://www.xs4all.nl/~timkr/chess2/diary_10.htm
http://www.xs4all.nl/~timkr/tour/breeze.htm
Both are pretty good reading.
Gambits should be fun. Chess should be fun. Enough of this stodgy "Oh, what if it's not sound" attitude.

Hm, yes, but if you say that something "should be fun" it's already close that it isn't. Yes I know that Halloween is called Mueller-Shultze, did these two men even exist ? I'd suggest that it's like Smyth and Jones.
The good gambits have ideas of position and initiative in exchange for some material, not of memorizing unsound lines until move 40.

I don't think Mueller and Shultze were actual people. But really, the opening is quite fun. Anytime you can trade advantages and play on the edge, it's fun. You mention the KG. I play a half dozen KGs a day, but I would rather play the double Muzio anytime than some drawish, mundane line in the KG. And the double Muzio is said to be unsound (but not actualy proven to be, that I know of).
I don't see anymore memorization with the Halloween than with any other gambit opening. Like any opening, there's a body of theory. I can see where white doesn't get any great advantage, with proper play from Black (a big if), but I don't see the concrete unsoundness - just noise, claiming it.

Maybe you are a brave batgirl then, I only play the gambits that won't leave me worse with correct play now. Well I play the Scotch and Smyth Morra but it seems to me that they are half sound, at least when I drink a bit.
Of course, an engine user can "refute" it.

http://www.xs4all.nl/~timkr/chess2/diary_10.htm
http://www.xs4all.nl/~timkr/tour/breeze.htm
Both are pretty good reading.
Krabbe discusses his first meeting with the Halloween Gambit in the "Breeze" article you cite.
I think that whether Halloween Gambit is unsound or not, i still can throw my opponent off his game and pretty much lock him up on his back ranks. I've played a tonne games in FICS using this gambit, and still no-one has refuted it the way batgirl did. Really a good job finding that refutation batgirl!

Is that supposed to be the refutation? I think Larry Kaufman recommended it years ago, but the recent results haven't been entirely convincing. As black, the masters who play the Halloween don't choose that line.

Krabbe discusses his first meeting with the Halloween Gambit in the "Breeze" article you cite.
I keep thinking there was an earlier posting at Chess Curiosities about this.

Is that supposed to be the refutation? I think Larry Kaufman recommended it years ago, but the recent results haven't been entirely convincing. As black, the masters who play the Halloween don't choose that line.
It's a "supposed" refutation. I'm not a theorist, but from what I can determine, it neutralizes the attack, and maybe makes normal gambit lines unusable.

It's a "supposed" refutation. I'm not a theorist, but from what I can determine, it neutralizes the attack, and maybe makes normal gambit lines unusable.
When I play the Halloween as black, I usually hold onto the piece, and try to force a queen exchange, and then win in the endgame. Kaufman's line looks good enough for equality, but if you are familiar with the theory, I think you can try for more.

Yeah, me too, but what I mean is I think it may be easier to win (or lose) if black tries to hold onto the piece. Giving back the piece seems to increase the probability of a draw.

Yeah, I think that's the nature of the supposed refutation - that it makes the game drawish. Maybe Black thinks the material superiority should be worth more than a draw and so avoids such a line? I love gambits that sac more than a pawn. I think the reason is that it does tend to make my opponent overly aggressive, even cocky, and also might make him feel that he must win.

Yeah, I think that's the nature of the supposed refutation - that it makes the game drawish. Maybe Black thinks the material superiority should be worth more than a draw and so avoids such a line? I love gambits that sac more than a pawn. I think the reason is that it does tend to make my opponent overly aggressive, even cocky, and also might make him feel that he must win.
Yeah, I agree completely. The Halloween in particular seems to affect people in a very strong way, especially if they've never faced it before.
Once queens are off, its an equal game.
You're forgetting that White's a pawn up.
A pawn up still might not be enough to bring home the full point with little tactical possiblities left on the board
That doesn't mean though that "it's an equal game" as you said.
and trading queens will develop black's bishop and then black can quickly pressure the d pawn and win it. But the way these guys play, anythings possible.
Presuming White doesn't self destruct (which he probably would, but obviously we have to presume he doesn't when analyzing), Black can't win the d-pawn.