Kling and Horwitz Defense

Questions arose while going over Ending 56 in Jesus de la Villa, 100 Endgames You Must Know: Vital Lessons for Every Chess Player, 4th ed. (2015), 129-132. Although de la Villa indicated that moving the king to the long side is an error while defending against a center pawn, his analysis showed that Black could hold. In one case, Black's king was on the long side while Black's rook harassed White's king from the short side. De la Villa was quick to point out that this plan fails against a c- or f-pawn, although it works against a center pawn. The technique, according to de la Villa is called the Kling and Horwitz defensive technique.

Curious, I dug into the newest endgame book on my shelf: Bernhard Horwitz and Josef Kling, Chess Studies and Endgames, updated and ed. by Carsten Hansen (2024). Quite a few endgame books were pulled from my shelf over the next few days as I studied the positions, analysis, and terminology.

First, I sought the original position in Kling and Horwitz. It was not a simple matter to find a position with any of the technique that de la Villa showed in 100 Endgames. I looked elsewhere for some clues to why the technique is dubbed "Kling and Horwitz".

Grigory Levenfish and Vasily Smyslov, Rook Endings, trans. Philip J. Booth (1971) credit Josef Kling and Bernhard Horwitz with the following position that is a win with White to move, but a draw if Black has the move.

Aleksandar Matanovic, Encyclopedia of Chess Endings, vol. 2 offers the same position flipped with Black's king on f7 and the other pieces on the d-file. It also is credited to Kling and Horwitz. Searching for the position in Kling and Horwitz, Chess Studies, or Endings of Games (1851) is a futile quest. The position appears no where in that book. Nor does Horwitz and Kling, Chess Studies and End-games (1889) offer it.

Sam Shankland, Theoretical Rook Endgames (2023) favors the term "long and short side defense", but does indicate it is also known as the Kling and Horwitz defense (27). Shankland also deploys the term, “botched Philidor defense” (27).

The position in Levenfish and Smyslov differs from one reached in analysis by Francois-Andre Philidor only in the significant placement of the defending king on the long-side of White's pieces. Philidor's analysis intended to show the pitfalls of the weaker side failing to maintain the rook on the sixth rank.

White to move

Philidor, Analyse de jeu des Echecs (1777)

1.e5 Rb6

Philidor asserts that if Black deserts the sixth rank with the rook before White pushes the pawn to the sixth rank, then White wins. In a variation, he attempts to show the necessity of the sixth-rank defense beginning with 1...Ra1. Philidor's variation bears some attention because, as has been shown by several writers, 1...Ra1 is not the error that Philidor thought. While the sixth-rank defense in his main line in simple and unbeatable, the positions in the variation have practical significance. It frequently occurs in play that a textbook Philidor position cannot be reached, and yet there are still ways to draw.

2.Kf6

Black to move

A position similar to this one is often presented in discussions of the Philidor position when it is the stronger side's turn to move. Black can no longer achieve a third- or sixth-rank defense because the pawn will block the check with an unstoppable mate threat.

However, Black can check from the rear. Philidor asserts that this defense fails if the stronger side plays it correctly.

2...Rf1+ 3.Ke6 Kf8

3...Kd8 does indeed lose.

4.Rh8 Kg7 5.Re8

According to Philidor, this is the only move that assures the win. In fact, as engines now confirm, the position has remained drawn the whole time.

5...Re1 6.Kd7

Black to move

6...Kf7??

No less than seven alternatives would have held the draw: Kg6, Re2, Re3, Re4, Rd1+, Rb1, Ra1.

7.e6+ Kg7

Philidor also offers 7...Kf6 8.Rf8+

8.Ke7??

After this error, Black can again hold the position. White’s rook needed to move from in front of the pawn to d8, c8, b8, or a8. The last seems most principled as it denies Black access to the a-file.

8...Re2??

From this point on, Philidor's analysis correctly shows how White wins.

8....Ra1 was necessary. With the king on the short side and the rook on the long side, Black is able to hold the draw either checking the king from the side or repositioning the rook behind the pawn to prevent its advance.

Jesus de la Villa explains the technique well using a position that he apparently composed.

Black to move

There may be worse places for Black's rook, but perhaps no worse ones where a draw is still possible. De la Villa writes, "I have deliberately chosen an unfortunate position for the black rook in order to illustrate all the problems that may arise" (129-130). He emphasizes that with a center pawn, placing the rook behind the pawn can hold a draw, "even if the defending king is forced to the long side" (129). In his analysis, Black's king is not forced to the long side, but rather moves there deliberately so that de la Villa can best illustrate the drawing technique even when there are difficulties.

Nine moves into his analysis, the position in Levenfish and Smyslov is reached.

1...Rf1 (Rf2, f3, and f4 all work, too) 2.Kd6 Re1

"This is the K&H 'trademark'. The rook stands behind the pawn to prevent its advance", according to de la Villa (130).

3.Ke6 Kd8 (Kf8 is the correct way to proceed, but this "error" is not critical).

4.Rh8+ Kc7 5.Kf6 Kd7

This important move was missed by an online opponent of mine a few months ago, allowing me to show my handling of the Lucena position.

6.Rh7+ Ke8

Returning to the promotion square should always be part of the weaker side's defensive strategy.

7.Ke6 Kd8 8.Rh8+ Kc7 9.Re8

"The only serious attempt to make progress" (De la Villa, 131). The position is similar to the one where Philidor wrote, "This is the only move which can assure you the game" (Analysis of the Game of Chess, 1777, 279).

Black to move

9...Rh1

Black prepares side checks from the short side.

10.Rf8 Re1

The rook returns to behind the pawn.

De la Villa's analysis ends here. Levenfish and Smyslov carry it one move further.

11.Rf2 Kd8

The position in Levenfish and Smyslov and after the ninth move in de la Villa appears in a variation in Dvoretsky's Endgame Manual (2003). Missing from Dvoretsky is credit to Kling and Horwitz. In fact, Dvoretsky gives the ideas without crediting any endgame theorists or composers from the past. With the 5th edition (2020), however, credit is given to Max Karstedt. I have not confirmed that the 3rd and 4th editions do not credit Karstedt because I do not have them at hand, but my hunch is that this bit of information represents Karsten Mūller's editorial contribution.

The Karstedt Maneuver 

Ilya Rabinovich, The Russian Endgame Handbook, published in the Soviet Union in 1938 credits Karstedt and does not mention Kling and Horwitz. Yuri Averbakh, Comprehensive Chess Endings, vol. 5 Rook Endings was first published in 1984 in Russian, then in 1987 in English translation. Averbakh writes about this position:

Black to move

"Philidor thought that ... Black could draw only by 1...Rb6, and that any attempt to attack from the rear--1...Rb1 would lose. But Karstedt showed that this is not so, and later analysis by Berger, Cheron and Rabinovich basically confirmed his conclusions" (123). Having spent some time studying Rabinovich's analysis before I found this statement by Averbakh, I created with some help from Stockfish the analysis of Philidor's discussion that I present above.

Inasmuch as Kling and Horwitz's work is readily available both the originals digitized and accessible through Google Books and through Carsten Hansen's new republication of the later text, which wholly absorbs the 1851 text, I am satisfied that the term, "Kling and Horwitz technique", is an error.

If I could locate Max Karstedt's original analysis, I could be certain that the "Karstedt maneuver" is the correct term. However, Kartsedt left us no books. He did compose about 150 studies. None of those available in Harold van der Heijden's Endgame Study Database show this technique. Karstedt is not well-known. There is a Wikipedia article about him that offers a references to a secondary publication: Siegbert Tarrasch's book on the 1906 Nuremburg tournament. A few copies are available, but not at a price I'm willing to pay. This secondary reference is Wikipedia’s source for an article by Karstedt that is said to have his appeared in "German Weekly Chess" (presumably Deutsches Wochenschach).*

A date of 1897 is given for Karstedt’s correction of Philidor in several texts, but reference to an original source appear in none of the chess books I checked. My best clue as to where to look comes from Siegfried Hornecker, "The Unknown Soldiers I", ChessBase News (30 May 2020). Hornecker cites Die Schwalbe (February 2018), where Günter Büsing states that Karstedt wrote for Deutsches Wochenschach and Deutsche Schachblätter. This makes the Wikipedia reference possibly accurate. Although some issues of both of these periodicals can be found online, the 1897 edition of Deutsches Wochenschach is not available, and the other periodical began in 1906. With access to the 1897 issue of Deutsches Wochenschach, I would expect to be able to locate an article by Karstedt. It would be an interesting read.

Lacking appropriate primary sources, I might choose to trust Rabinovich, Averbakh, and Müller. Alternately, I might consider the term Karstedt maneuver speculative.

Whatever name is employed, it is worth learning. For that, Jesus de la Villa and Sam Shankland both offer exceptional in-depth analysis.

*Wikipedia shuns any work that appears to be “original research”, and hence has a built-in preference for a secondary work that mentions Karstedt, such as Tarrasch, over anything Karstedt may have written. I’m not certain whether there is a preference for translated names of publications, or whether this is simply sloppy editing.