MNb wrote on 08/12/11 at 20:34:12:

What counts in chess are concrete moves. I gave analyses showing that 1.e4 c5 2.Nc3 e6 3.g3 (or even 3.f4 d5 4.d3) 3...d5 is good for White. Then you claimed again that 2.Nc3 wasn't to be recommended, and that, somehow, 2.Nf3 was the right move. While I believe that many "Antis" are fine weapons (including b3, which I played for a while), I am not religious about converting you to give up 2.Nf3. Actually I like this move very much. In my mag I published an article "proving" that 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 e5 6.Ndb5 d6 7.Nd5 Nxd5 8.exd5 Nb8 9.Qf3!? is advantageous for White. However, months later in a strong corr. game an equalizing line was found which I was unable to repair. 

So what is your antidote against the Sveshnikov, MNb? There are other defences which are perfectly sound for Black, but the forcing character of the Sveshnikov makes it particularly hard to find a promising line for White. - Many express their deep belief that 2.Nf3 is the mainline, without ever doing much analysis of their own. They "have heard" that 2.Nf3 is the main line, or they admire top players who are playing 2.Nf3, but sorry, I am not impressed by such arguments. In the 1860s they thought 2.Bc4 is clearly the best move. Today we are still in the stone age of theory, so who can really be sure which move is best? 

Apparently you misunderstood my tipping point argument. Is there really no point - 80%, 90%, 99% - at which you say: "it's too much - now I am playing 2.Nc3"? Psychology is part of the game, and if the overwhelming majority of players - perhaps with the exception of Schaakhamster's town - prefers 2.Nf3, we might want to inquire the alternatives. You cannot really think that 2.Nf3 has all (open positions, dynamics, novel ideas, winning potential) and 2.Nc3 nothing of it. 

Concrete moves are king, but sometimes I try to make the value of alternatives plausible with general observations. Here is my article about 2.a4: http://www.chesscafe.com/text/kaiss65.pdf . In my opinion, Myers' Variation 1.e4 c5 2.a4 is a refined version of the Rossolimo. Objectively 2.Nc3 may be better, but its theoretical body has become relatively large, too. In my suggested repertoire for 1.e4 players therefore I decided to prefer 2.a4. - 2200 Levenfish vs perhaps five important 2.a4 games, still a small difference. 

Stefan Buecker wrote on 08/12/11 at 11:52:56:


The direction in which these souls will be heading after moves like 2.a4, I would not associate with salvation.

I'm keen on seeing a good analysis on 1.e4 c5 2.b3 (or 1.e4 c5 2.Nf3 ... 3.b3) with all the good ideas that white can produce after.. 

Stefan Buecker wrote on 08/12/11 at 11:52:56:


Well that is not my experience as an average chess player. Perhaps it is more the case at higher levels but if you play the sicilian around here you'll be hit by a barrage of anti-sicilian (all sorts...). 

When you see how many people after 1.e4 c5 play 2.Nf3, and that only a minority plays differently, it isn't surprising that in a forum like this the "Play the Open" fraction is twice as strong, or more, and the "Antis" are only a choir of dissonant voices (2.c3 or 2.b3 or GPA or...). Often the 2.Nf3 fans are also the stronger players, so the forum has a strong bias. 

I believe that in our situation today, with 70% playing 2.Nf3 followed by 3.d4, the alternatives are underestimated and not sufficiently prepared by the Blacks. It can be lucrative to profit from the prejudice and study a rarer second move in depth. But it is only my intuitive feeling, statistics cannot really prove my case. The reasoning above that there must be a point when 2.Nf3 is played unreasonably often is probably an argument most people will agree with. But at which percentage do we reach the tipping point? To this key question I don't have an answer, nobody has. It is a bit like the "Laffer curve", and we know how utterly useless in practice this theoretical idea was. So I am writing my articles, on 2.Nc3, 2.Bc4, 2.Na3, 2.b4, 2.a4. That's all I can do to save at least a few souls.

I recall reading in a few sources that White's winning percentage increases from 51-52 to around 56-57 as you reach the highest levels.  In particular it is hard for Black to play for a win at the highest level if White is happy with a draw.

On the subject of losing openings, I think the range may well actually be quite narrow, unless we limit the search to forcing sidelines (such as 1.e4 e5 2.Nf3 f6 3.Nxe5 fxe5).  I reckon that the Halloween Gambit might just give White the worse of a draw (Black's advantage was =+ according to Wind in Kaissiber and might stretch to -/+, but it isn't decisive).   Similarly, the Omega Gambit (1.d4 Nf6 2.e4) seems to give Black an advantage in the region of =+ to -/+, so I think White can probably draw there with best play.  That's even before we get around to the more respectable gambits like the Morra and Blackmar-Diemer mentioned above, which are probably equal (some analysts try to claim a small edge for Black, but nowhere near winning).

Considering some maligned Black openings, having been involved in extensive analysis of the Englund Gambit, I believe Black probably has the worse of a draw there, as many of the deepest lines leave Black with good drawing chances in a worse ending, and the Latvian is much the same.  Even after 1.e4 f5 2.exf5 Nf6 I would merely assess the position as clearly better for White. 
 
On the other hand I reckon that the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+) is probably lost for White with best play.

MNb wrote on 08/10/11 at 16:04:04:

With 1.e4 c5 2.Nc3 e6 3.g3 d5, for players 900-1800 Elo, year 1975 or later, I find 514 games in MegaBase 2008, the result is 48% for White. (2500+: 58%)

These 2.a4 games (133 games) were not from MegaBase alone, many engine games - so the 57% are irrelevant. From the eight games in the MegaBase, only the two games by Gerard Welling showed an understanding of the position. Still, a player who hates the gambling of opposite castling, who knows his "Closed" well (and there are subtleties in 1.e4 c5 2.Nc3 e6 3.g3 d5, as in 2.a4 or in the Rossolimo), isn't always well advised to play the Open Sicilian.

Stefan Buecker wrote on 08/10/11 at 07:37:58:


The question is, does White score 57 % because of 2.a4 or despite of? As we have discussed before, openings rarely decided the outcome in games between U1800 players. But of course it helps to be able to choose how to start the game. It seems to me that often in these games there is a race to deviate first in order to avoid the opponents opening knowledge in the main lines. Hence the many Exchange Slav, QGD, Caro-Kann, French, Spanish etc. These are the first lines I prepare for when learning a new opening, and the first I look at when preparing for a game.

Even under 2200 there are many players that don't want to play the main lines out of fear of getting beaten by book moves.

This could be a reason "objectively" unsound openings do reasonably well, because you don't prepare for these because you skip over them due to time constraint and ignorance. The people that play these openings still know how to handle most of the positions and U1800 this is believed to be more critical than playing a sound opening I think. However, I would rather try to get a good positions based on sound moves than take a chance with an unsound variation.