I can add one small thing about not knowing theory. While there are more theortical information available than ever before players  bellow elite leve know less and less concrete theory. I have met a several FM:s who has been out of book before move 7 in mainlines regardless of side.
If have only met 3 GMs but in each case I knew theory some move longer.

There are three kinds of lies: lies, damned lies, and statistics. Statistics does not always show the correct evaluation of a line.
I can give an example from an other opening with equal material that several books for black recommened where all engines thinks dead equal while in practical  play white scores 61% in databases. It could be that in practical over the board games whites position is easier to handle when one is out of book.

ErictheRed wrote on 08/12/17 at 17:47:38:


I take your generalization about strong players not necessarily knowing (or remembering) theory which is certainly plausible in some cases.  I still recall the post you made many years ago (2010?) where you shared how quickly you were "out of book" in your previous 12 OTB games.  However, that wasn't really the main point of my post.   

Most of the benefit I derived from this was from personally analyzing every single one of the losses and understanding where Black went astray.  Since I can't post all the details, an interesting summary observation for me was that Black was most often the one varying from main line theory rather than White and in many cases that put them under pressure in an inferior position.  Generally I would expect it to be the opposite since French players tend to be specialists and the main lines in the 3...Nf6 Tarrasch are somewhat "narrow".   

Another helpful observation was that I couldn't find anything  theoretically suspect about the 3...Nf6 line even though it is not as popular or successful lately.  This was personally encouraging for me since I find the popular 3...c5 lines to be more drawish than I like.  My view is that the main ideal of opening knowledge is to get you to a playable middlegame which you like and hopefully understand a little better than your opponent!

A few months ago I did an analysis of recent trends in the Tarrasch and thought I'd share some summary observations since it's been quiet on the forum.  I looked at which variations for Black versus the French Tarrasch are currently (2016-Mar 2017) most popular among strong players (rated 2400+) and how the various tries for Black holding up in competition?   
I looked at 533 games from the position after 1.e4 e6 2.d4 d5 3.Nd2 that were played between January 2016 and March 2017 where both player’s rating was greater than 2400.

Initial Observations:
•      Overall White scored an impressive 57.5% with the French Tarrasch winning 179 games, drawing 255 games and losing 99 games in my sample.
•      This compares favorably with White’s overall performance after 1.e4: 53.2% or versus the French in general 1.e4 e6: 53.8% from that same database.  Indeed the Tarrasch was the highest scoring variation for White (including all ratings) even as compared to the more popular 3.Nc3.
  •      The 3…c5 variation has eclipsed all others in popularity for strong players of the Black pieces.  For the sample, 48.8% of Black players rated 2400+ chose 3…c5 versus similarly strong opponents.  This was followed by 3…Be7 (16.3%), 3…Nf6 (12.4%) and 3…dxe5 (8.1%).  Somewhat surprisingly the Guimard 3..Nc6 came in fifth at 7.3% followed by 3…a6 (3.3%) and the eccentric but lovable 3…h6 (2.4%).

  The 3…Nf6 line has been a perennial favorite receiving recommendations from John Watson in PTF4 and Emanuel Berg in GM16.  I was curious why it has slipped to third in popularity and is scoring the most poorly (aside from the Guimard)? 
I compared the losses suffered by Black to the recommendations in Berg’s repertoire to see where the games varied.  After looking at the 28 losses by Black in the 3…Nf6 line, it is interesting to note that:
•      Black chooses to vary from the recommendations in Berg’s repertoire by move 11 in 82% (23/28) of the games.   
•      Only one White win, Artemiev-Onischuk follows Berg’s “absolute main line” after 5.Bd3 and Black varies by failing to execute the thematic French exchange sacrifice 16…Rxf3 and suffers with an inferior position.
•      It seems as though Black frequently tries rare sidelines in order to surprise White but this contributes to his downfall.

In summary, it appears that Black is often the cause of his own undoing by deviating from the recommended 3...Nf6 lines.   The variation as a whole remains sound and would score much better for Black if these decisions were avoided.

I hope this is interesting / helpful to the forum members and stimulates more discussion.  This forum was particularly influential when I returned to OTB chess a number of years ago so I'll always be grateful to the members, administrators and Tony for making it possible.