He Refused To Develop a Single Piece For 17 Moves

There are certain opening principles that we learn almost as soon as we start playing chess.
Fight for the centre. Castle your king. And, perhaps most importantly, develop your pieces.
Emil Joseph Diemer apparently had some different ideas.
In this game from the 1984 Nuremberg Open, White played 17 pawn moves in a row before moving a single piece. No knights, no bishops, not even the queen. For the first 17 moves, every single White move was made by a pawn.
And yet, that isn't even the strangest part.
If somebody deliberately wanted to set a record for the most consecutive pawn moves, doing something like this wouldn't be particularly interesting. You could simply push random pawns while making your position more and more terrible with each move.
But that isn't what happened here.
Most of Diemer's moves were perfectly legitimate. Some weren't objectively the best, and a few were certainly questionable, but there was nothing resembling obvious trolling. There was usually a logical positional or tactical reasoning behind what he was doing.
Somehow, the position simply kept presenting White with another reasonable pawn move.
And another.
And another.
And then we eventually arrived here.

17 Pawn Moves In A Row

What Makes This Game So Remarkable?

There probably isn't some profound chess lesson to take from the rest of this game. There was a noticeable strength gap between the players, Black made several serious mistakes, and once Diemer took control, the conversion was relatively straightforward.
But I still think the opening makes this one of the strangest games I've ever come across.
Playing 17 pawn moves in a row isn't even necessarily that impressive. If you put that down as your goal before the game starts, you could probably manufacture something similar, especially playing against a much weaker opponent.
But that's just not what has happened here.
Look back through Diemer's moves. Yes, there are inaccuracies, for sure. 5.g5, 8.c3 and particularly 17.f5 are questionable. But none of them look like somebody deliberately ruining their position just to achieve some arbitrary goal of playing as many pawn moves as possible.
There is a recognizable idea behind every move.
g4-g5 grabs kingside space. f4 builds a huge central trio of pawns. d5 closes the position. c3 challenges b5. h4-h5-h6 gains space and forces Black's bishop backwards. a4-a5 attacks Black's queenside and drives the knight away. exd5 restores the central pawn. c4-cxb5 attacks Black's queenside structure. Even 17.f5, probably the worst of the sequence, is still an understandable (from a human standpoint) attacking move.
These really aren't completely random pawn pushes.
And I think the closed nature of the position was crucial. Black never opened the centre quickly enough to punish White's complete lack of development, which allowed Diemer to keep finding another pawn move without his position collapsing. The exact circumstances required for something like this to happen naturally are incredibly unusual, which is why a game like this is so rare.
Perhaps the most convincing detail comes immediately after the streak ends.
After 17...gxf5, White could have played 18.b6, extending the sequence to 18 consecutive pawn moves. It wasn't some ridiculous move either. Yet Diemer instead played the stronger 18.Qh5+.
That makes me think the record itself wasn't really the point.
Diemer didn't play 17 pawn moves because he was desperately trying to play 17 pawn moves. He just kept finding legitimate reasons to move another pawn, until after seventeen moves the position finally demanded something else.
And that, more than the number itself, is what makes this game so extraordinary.
Seventeen moves. Seventeen pawn moves. Zero pieces developed. And somehow, it was still a perfectly serious game of chess.