I've just been having a look at the Steinitz Gambit and I independently reached the same conclusion.  The immediate 10.a4 limits Black's options (e.g. 10...f5 is no longer as strong because of 11.a5).  I think 10...Bg7 is indeed best, with the idea 11.a5 Nge7, planning to castle kingside.  Thus White should play 11.Nd5, encouraging 11...0-0-0.  My analysis continued 12.a5 b5 13.Qe2 a6, shoring up the queenside, and threatening ...g5-g4, and Black is slightly better, but White has some attacking chances on the queenside with the idea of c2-c4.

http://www.chesspub.com/cgi-bin/yabb2/YaBB.pl?num=1217111507/47
I don't have anything to add to the analysis of the Hamppe-Allgaier as while White may or may not have marginal improvements over Brabo's analysis, it's clear that White's compensation is probably insufficient for full equality in the 8...f3 line.  But it may not matter as 5.g3!? is currently getting favourable attention, even at Grandmaster level, via the Quaade Gambit (when 3.Nf3 g5 4.Nc3 Nc6 transposes).

Thus, another reason for preferring 3...Qh4+ 4.Ke2 Ne7! =+.

SWJediknight wrote on 11/06/10 at 19:23:48:


1. e4 e5 2. f4 exf4 3. Nc3 Nc6 4. Nf3 g5 5. h4 g4 6. Ng5 h6 7. Nxf7 Kxf7 8. d4 f3 9.gxf3 Be7 10.Be3 (9.Be3 Be7 10.gxf3 is a simple transposition) d6 11.Bc4+

A) 11....Kg7 12.Bf2 Nf6 13.Qd2 d5 14.exd5 Na5 15.Bd3 Nxd5 (All this was already published by myself at 25th of October) 16.fxg4 Bb4 and I can't see sufficient compensation for the sacrificed piece

B) 11...Ke8 12.h5 Nf6 13.Qe2 Rf8 (gxf3 should also be good for a serious advantage for black) 14.e5 dxe5 15.Qd3 e4! 16.fxe4 Nb4 17.Qd2 g3! and the complications seem to be winning for black.

I believe in an OTB game you can get away with this but in a high level CC game I think you will suffer with white.

Could someone please correct the title of this thread?

As far as I know this opening is known as the "Mason gambit", Masion looks like a misspelling of "maison" 

SWJediknight wrote on 11/06/10 at 19:23:48:


After 8...f3 I prefer an immediate 9.Be3.  At worst, it's a transpositional device that leads back into the lines under discussion.  At best, it encourages 9...fxg2 which I believe to be winning for white based on two correspondence games I played in 2009 (included below).  Having said that, 9...d5 and 9...Bb4 also deserve consideration but that's for a separate post...

I have much more to add to this discussion but before I do that I'm going to go back and try to organize all of the analysis posted thus far into PGN format, so that I can be sure that I've not overlooked an important contribution.  

[Event "?"]
[Site "Online Chess"]
[Date "2009.09.21"]
[Round "?"]
[White "R. Gates"]
[Black "mortzy"]
[Result "1-0"]
[PlyCount "71"]
[EventDate "2009.??.??"]
[TimeControl "1"]

1. e4 e5 2. f4 exf4 3. Nf3 g5 4. Nc3 d6 5. h4 g4 6. Ng5 h6 7. Nxf7 Kxf7 8. d4
f3 9. Be3 fxg2 10. Bxg2 Be7 11. O-O+ Kg7 12. e5 c6 13. Qe2 d5 (13... dxe5 14.
dxe5 Be6 15. Be4 Qc8 16. Rf4 h5 17. Raf1 Rh6 18. Qf2 Qe8 19. Bf5 Kh8 20. Bxe6
Rxe6 21. Rf7 Bf6 22. Rxf6 Nxf6 23. exf6 Qg6 24. Ne2 g3 25. Qf3 Qg4 26. Nf4 g2
27. Rf2 Qxf3 28. Rxf3 Rd6 29. Ng6+ Kh7 30. Nf8+ Kh8 31. Bc5 Rd1+ 32. Kxg2 Na6
33. Be7 Nc7 34. Rf5 Rd5 35. Ng6+ Kh7 36. Ne5 Ne6 37. Rxh5+ Kg8 38. Kf3 Rd1 39.
Kg4 Rf1 40. Rf5 Rg1+ 41. Kh5 Kh7 42. Nf3 Rh1 43. Ng5+ Nxg5 44. Rxg5 b6 45. c4
a5 46. Rf5 Rh8 47. f7 Kg7+ 48. Kg4 Rg1+ 49. Kh3 Rf8 50. Bxf8+ Kxf8 51. b3 {1-0
R. Gates-qazzaqy2k/Online Chess 2009}) 14. Nxd5 cxd5 15. c4 Rh7 16. cxd5 Bxh4
17. Be4 Rh8 18. Qh2 g3 19. Qh1 Bg5 20. Bxg5 Qxg5 21. Qf3 Nf6 22. exf6+ Kf7 23.
Rae1 Qh4 24. Re2 Re8 25. Bg6+ Kxg6 26. Qd3+ Re4 27. Qxe4+ Qxe4 28. Rxe4 Bh3 29.
f7 Nd7 30. Re8 Bxf1 31. Kxf1 Kxf7 32. Rxa8 a6 33. Ra7 Nf6 34. Rxb7+ Ke8 35. Rb6
Nxd5 36. Rxh6 1-0

Btw, I think this opening exposes the limitations of computers better than just about any other.  The primary game above is an especially nice example (if I do say so myself!)  The computer thinks white has gone mad after the second piece sacrifice but by that point black is nearly busted!


 

Quote:

Yes, Rybka does seem to have some failings of its own- e.g. the failure to recognise elementary king & rook's pawn vs king as a draw (whereas even Crafty and Fritz 5.32 recognise it as drawn straightaway) and I remember seeing a Rybka 3 analysis of Blackmar-Diemer Gambit lines where it saw "best play" as being the loss of a tempo with Bb5+, c6, Bc4.  The 2.3.2 version also doesn't seem to find forced mates as quickly as Fritz 10.

I'm not really a die-hard computer analyst (or for that matter a Mason or Vienna Gambit practicioner, the latter arising via transposition if Black plays 3...Nc6, though I do occasionally face these lines with Black)- this is really an academic interest in chaotic gambit lines, and I only really tend to update my commercial chess applications every few years or so.  I also think there are certain situations where no computer really handles "compensation for sacrificed material" correctly, so I don't always trust their assessments.

I think White could feasibly have a lot of fun with the Mason Gambit outside of correspondence play and high-level OTB play, judging by the chaotic nature of the resulting positions, but have serious doubts about its viability (other than as a surprise weapon) otherwise.   I think the same is probably true of the Vienna Gambit and associated Hamppe-Allgaier/4.d4 lines that were discussed earlier (I found no clear route to equality for White against your 8...f3 by the way), though I maintain the view that Black has to tread a narrower tightrope in those lines in order to hope for an advantage.

TalJechin wrote on 10/30/10 at 22:54:47:

As the game develops opportunties will arise, yes sure but I thought we try here to search for the truth. If you search for a playable position for OTB then almost anything is ok. Just have a look at the games of Magnus Carlsen and you will see that he often gets away with a lot of shaky opening experiments even against worldclass players in official games.

1. e4 e5 2. f4 exf4 3. Nc3 Nc6 4. d4 Qh4+ 5. Ke2 b6 6. Qd2 Ba6+ 7. Kd1 Bxf1 8.Nf3 Qh5 9. Rxf1 g5 10. Nd5 O-O-O 11.a4!? (Interesting but I don't believe that this new move changes the evaluation.) f5 12.exf5 Qf7 13.Nc3 Qxf5 14.a5 b5 15.d5 b4 (15... h6 16.Qe2 Nb4 17.Nd4 Qg6! must also be better for black because now 18.Ndxb5 fails on Re8 which wasn't possible with 17... Qf6. Of course there is still 18.Qxb5 but the complications after c6 favour black. I find b4 more professional, not trying to win now but being satisfied with a nice endgameadvantage which arises after ...) 16.Qd3 Qxd3 17.cxd3 Ne7 and I estimate the chances 50/50 between draw and a black win.