I've been absolutely destroyed by the Muzio. Granted, I'm not that good a player, but if you have any doubts, try playing someone who studies it.

Call me crazy (or blinded by love?) but I actually think the Halloween Gambit holds up well in correspondence chess unless black happens to be an aficianado of the gambit in which case white's playing for a draw. 

Most of the Hallowee theory that's publicly available is at least partially outdated and none of the popular opening books contain anything close to the latest theory since they're all so Playchess-centric.  If that wasn't bad enough, few of the landmark Halloween games are in the various online game repositories or even the Chessbase Mega DB.  Unless you're an avid reader of Kaissiber or a UCO newsgroup follower, it can be difficult to properly prepare for -- at a minimum, it's much more work than *most* players are willing to put in, especially for an opening that looks so damn unsound....

As long as we're changing the subject to talk about other lines in which white gambits a whole knight, how about the Pierce Gambit? It's like the Muzio, but it comes from the Vienna Gambit - 1. e4 e5 2. Nc3 Nc6 3. f4 exf4 4. Nf3 g5 5. Bc4 g4 6. d4 gxf3 7. O-O and only then does white start grabbing pawns on the f file.

See these articles:

http://www.chesscafe.com/text/kibitz96.pdf

http://www.chesscafe.com/text/kibitz97.pdf

So many gambits, so little time! 

One thing that immediately caught my eye in the Harding articles was the following "...the Pierce gambit can probably only be refuted by accepting it, and White may well stand better in all other variations".  That's such an important consideration when trying to decide whether or not you want to learn a new gambit.  The modern game has become so cautious that many players immediately look for ways to decline gambited material in an effort to avoid irrational positions at all costs.  What's the point of learning the intricacies of a new gambit if it will only be accepted 1 in 10 tries?  

Thanks for pointing this one out -- it's on my "someday maybe" to-do-list.

Roy

maybe you haven't played a really experienced halloween player..
i've played the halloween gambit in nearly every game. currently, 60 wins 0 loss 0 draws with it. anyone wanna try and prove its unsoundness? :P

here is the game folks:

No, but I'd love to see those games. Apparently, there used to be a huge database of games on the internet in pgn format for the Halloween Gambit, but the link to it is broken now. It included a lot of games by a computer program that used to play the gambit at every opportunity against opponents on ICC, and it apparently beat up on quite a few masters in blitz games. I would think that going through the games in that database would be the best way to learn how to play the gambit.