It is possible that <goodevans>'s post meant to refer to <mutual> zugzwang positions (sometimes referred to by the term "trebuchet", especially in pawn endings; see: https://www.chess.com/chessopedia/v...), but if one considers the position after <39. a4> with Black's h-pawn removed, Black would not be in zugzwang. He would still have a waiting move with his king, such as 39. ... Kf6, which would leave White in zugzwang.

click for larger view

<29.Ng4> would be aggravating, since Black cannot protect the e-pawn (29...Rc5 30.Nxe5!). And this holds for several moves, becoming especially bad after <31...Kf6>

click for larger view

When 32.Ng4+ simply wins the e-pawn directly and probably the game as well. (OK, maybe not against Capablanca, but still...)

When two strong players -- and Marder wasn't all that bad -- miss a simple continuation, it can be shrugged off. Time pressure, distraction, fatigue, or any of a dozen different reasons. When both strong players not only miss a simple continuation, but do so for several mvoes and even make the situation worse by their play, the reason is almost surely a bad score.

Now, I've looked into this tournament a bit (something I recommend to anyone who values their insanity) and had a few of the early sources of the game at hand. Here's the earlies I found, from the <New York Sun>, August 3, 1913:

http://chroniclingamerica.loc.gov/l...

If you don't want to bother, I'll tell you -- it's identical to our score. So if a mistake is in there, it's been there form the beginning and has simply been copied by everybody else. You don't need a computer to do that.

Why wasn't this brought up then? Because players like Capablanca get a pass on this sort of thing. We assume there's some devilish trap and move on.

After all this, here's my suggestion: after <28.Kg2>

click for larger view

28...P-QR4 is a typo for 28...P-KR4. Simple, easy, uncomplicated, and makes perfect sense as the game goes on. And one added benefit: after <39.a4>:

click for larger view

It turns out the quiet killer move is <39...a5!> -- even better than the game finish.

That is why I say that although this position is very similar to a zugzwang it isn't truly a zugzwang.

THe discussion begins here:

I'm delighted that you took such an interest in my comment and glad that together we uncovered the truth - although the vast majority of the credit of course belongs to you. Great detective work!