For better of worse, this is what I think about the position: Black pawns initially interfere with the mobility of both bishops. It is thus imperative to increase the scope of the Black bishop and not of the White one. Tarrasch's <39...c4?> has exactly the wrong effect; Black should have played <39...f4!> (IMKO)

I see this as the key drawing line: <39...f4! 40.f3! Bc6! ...>, eg, 41.fxe4 Bxe4 42.c3 f3 43.gxf3 Bxf3 ..., or 41.Bg5 exf3 42.gxf3 Bxf3 43.Bxf4 ...; both obvious draws.

That may seem a bit funny line and a bit strong statement; so let me expand a bit. The 40.f3! is actually a quite cunning move because the f4-pawn is doomed, in most variations, whether after exf3 or after the tempting e3. For instance, the game could go 39....f4 40.f3 Bg6? 41.c3 exf3 42.gxf3 Bb1 43.a3 Kc6 44.Bg5 Kd5 45.Bxf4 Kc4 46.Bc7 b5 47.Bd6 ... and White, with two extra pawns, should score. Or the game cound go 39....f4 40.f3 Bg6? 41.c3 e3 42.Kg1 Bb1 43.a3 Kc6 44.Bg5 Kd5 45.Bxf4 e2 46.Kf2 Bd3 47.g4 and, again, White is two pawns up and his winning chances are good.

Alternatives to 40.f3, such as 40.Bg5 and 40.Bd6, pose no problems for Black. For instance 39...f4 40.Bg5 f3 41.gxf3 exf3 42.Kg2 is a dead draw, and 39...f4 40.Bd6? is a tactical error as it allows 40...e3 and sudenly White must thread carefully not to loose the game by 41.Bxf4? exf2 or something like that.

"Besides the queen's wing majority, which eventually means a passed pawn there, White has an advantage on the other side because his king can attack Black's weak pawns. White has threats on both wings; and the king's side threat means either that Black's king will be drawn away from the queen's side, or that White will get a second passed pawn on the king's side –else it means nothing; for a pawn majority on one wing cannot win unless there are also threats on the other wing. Here Black would draw if his king's side pawns could be defended by the bishop alone."

For that reason, 39...c4? was a mistake, as <Gypsy> correctly stated. As he also mentioned, not 39...f4? because White gets a king's side majority by 40.♗g5 f3 41.g4. According to Euwe and Hooper, the correct continuation was 39…♗b5! 40.♔g3 (40.♔g1 ♗e2) ♗f1 41.h4 h5 42.♔f4 ♗xg2 43.♔xf5 ♗f3, and the black king blocks the pawn majority, while the bishop guards the pawns on the other side of the board.

This is a typical misinterpretation of Tarrasch's views on hypermodernism. In his book "The Game of Chess" Tarrasch makes a clear distinction between the practical value and the theoretical value of an opening.

The hypermodern openings, according to Tarrasch, have a practical value and will result in a draw with best play. However, and this is his main thesis, their theoretical value is inferior to the classical openings. For example, 1.e3 might be a good tactic over the board, and cannot be refuted by Black, but it does not have the same theoretical value as 1.e4 or 1.d4 which seize the center, maintain White's initiative, and aid in the development of his pieces. It's like an engine eval of .02 vs .08

Tarrasch's problem with hypermodernists like Reti and Nimzowitsch was that they not only maintained the practical equality of the hypermodern openings but also their theoretical equality as well, something Tarrasch could not accept.