<15.Qd2> The point where Duras steps off the theoretical paths of 15.Qe1.

<19.c4 Rd8 20.Nf3 Nf5> Superficially, Black position looks splendid, White dubious. But the inconspicuous passed pawn d3 gives White some chances. It seems that Reti misunderstands that and commences a way-too-optimistic plan.

<21...Nxg3+?> Of course, 21...Nd4, 21...Kf8 or even 20...Kf8 would have kept game in a drawish equilibrium. It seems that Reti hopes for more; but, because of the d3 pawn, White N proves superior to Black B at the end. The doubled White pans never become a liability in the game: Black has no sufficient piece play against them and it takes the same long time to break through them with Black pawns as it would take to break down a g+h pawn arrangement.

<22...Rc5(?)> Reti still does not understand the position and believes that he will get some pressure going against the g2+g3 pawns. But he just whittles away time; Kf8 was a good move to play.

<29.Ke3 Bb7(?)> By this time, White is significantly better: he has centralized K, passed pawn on d4, and a good coordination of his piecess. In contrast, Black K and Ra5 are mostly out of play. What a difference from the position of 10 moves ago! For better of worse, at this point Reti should have swaped his B for White N and search for salvation in a 4R endgame. I fully understand Reti's reluctance to part with such a nice-looking bishop. But that bishop is handsome on surface only.

<30.d5(!) ... 33.Kd4> Black bishop is now quite short of scope and White controls the center of the board.

<38.Kb4 Bd3> Because of the rather open position, one would think that the bishop should better the knight. But it does not, as the key lies in the position of the kings and in the d-pawn. As it is, the B is still severely dominated by the combined White forces. Btw, 38...Be8 was likely a better place for the bishop, but it seems doubtful that the game can be saved; the key problem then becomes how to save Black's a-pawn.