MNb wrote on 09/05/11 at 15:34:31:


Agreed, this is why I would be surprised if the Smith-Morra was ever refuted.  It is significant that all the gambits SWJediKnight mentioned as being in trouble theoretically are all Black gambits.

trw wrote on 09/02/11 at 14:10:30:

trw wrote on 09/02/11 at 22:18:22:

MNb wrote on 09/02/11 at 20:49:43:

Have you seen Esserman-Van Wely? My compatriot chose a defence many, including me, think highly of. White's idea Nd5 has been discussed on this forum. Despite Van Wely being 200 ELO-points stronger he got literally creamed in good old 19th Century fashion. Of course this says exactly nothing about the theoretical value of the Morra Gambit. It does confirm my point that it's not so easy to refute though.
I am more suspicious of the BDG because of 4...c6;/5...c6; but accept this is the only real attempt to refute it. As far as I can see everything depends on GM Gutman's idea as presented in Over the Horizons. So I can't go further than the stereotyped "awaits practical tests".
I have noted another tendency. It has been remarked many times before, I am not at all original. It seems that White's extra tempo allows to sac a pawn at a very early stage. At the other hand Black needs White to make a concession first (like f2-f4 of the GPA in Tal's Gambit and c2-c3 in the Marshall) to justify gambit play. If this is correct it should tell us something about the meaning of White's birthright.

Even if we had the solution of chess accessible to us, a very large amount of openings go beyond the horizon of human memory fairly quickly and remain unclear. 

What I'm glad for is that humans did the vast majority of the work first, before computers came along. In the future, like past the year 2100, I think people will still consider the vast majority of opening theory to have been done in the 20th and early 21st century. 
 

Another problem is that in some of the sounder gambits, when subjected to heavy computer analysis (with computers being very good at tactics) lines that offer double-edged and equal play increasingly end up analysed out to drawish endings (some lines of the Marshall Attack in the Lopez being stark examples).  It is often possible to research into sidelines that also offer interesting and equal play, but as computer-assisted theory advances the supply of such sidelines dwindles with time.

While I wouldn't call the Latvian, Englund and Elephant Gambits popular, until quite recently there were no clear-cut ways found for White to get a +/- against them, and thus they were quite viable at CC- but not now, as all three of them have been extensively analysed out to +/-.  The Milner-Barry Gambit against the French (with 9.Nxd4 rather than the modern 9.Nbd2) would be a good example of such a line for White.

On the other hand, I believe that computer-assisted analysis of the Morra, Wing(Sicilian) and Blackmar-Diemer gambits has if anything shown them to be more viable than was generally thought earlier, albeit only enough for equality.

trw wrote on 09/02/11 at 14:10:30:

That seems to be confirmed by White's provisional score in 2011 for ELO 2200+ - until you look at the expected performance.
The Morra Gambit has huge practical objections, but a theoretical refutation is not that easy. The Latvian and even the Jänisch-Schliemann seem to be in worse shape acc. do Madame la Theorie.
At the other hand it's a fact that strong computers have removed many gambits from the repertoires of strong corr. players. Fortunately for me there are a few left.

I think it's interesting that despite the huge impact the engines have had, very few openings have actually been refuted.  I think a few years ago a lot of gambits were expected to have been put out of business by now, but that doesn't seem to have happened.  Did we overestimate the power of the silicon beast?

Correspondence is a great test of this point of view, have people generally found this to be the case in practice, and can anyone give any examples of popular systems that are no longer viable?

Smyslov_Fan wrote on 08/24/11 at 16:05:12:

To each their own. But there is something satisfying to me to know I took out a GM in their petline where they had access to engines, databases and tablebases.

Smyslov_Fan wrote on 08/24/11 at 16:05:12:


This is why I said no real CC disallows computers. Joke sites like chess.com do.