Lucena/Philidor Positions in Silman's Endgame Book

[dennisk]

I have found Jeremy Silman’s Complete Endgame Course: from Beginner to Master to be the most useful endgame book for me. I regularly go back to this book to review the basic endings so that I don’t forget how to play them. However, a few years back, I discovered what I believe to be a mistake in one of the problem solutions. But before I (a humble C player) accuse IM and respected author Silman of an error, I thought I would get some second opinions.

For those who have the book, it’s Problem 9 in Part Four (Endgames for Class “C”). Silman’s solution is given on page 153. I don’t know how to embed a diagram, but it’s a simple position: White has King on d5, Rook on a2, and Pawn on e5; Black has King on d7 and Rook on g8. It’s Black to move. Silman’s statement of the problem: “White threatens the unpleasant Ra7+. How can Black defend himself?” His solution follows:

If you thought you could draw by trading Rooks, you’re wrong: 1…Rc8 2.Ra7+ Rc7?? 3.Rxc7+ Kxc7 4.Ke6! followed by Kf7 is an easy win for White. Instead, Black can effortlessly hold a half point by using the classic Philidor formula.

1…Rg6! 2.Ra7+ Ke8

Now White’s King can’t advance onto the 6th rank. Since Black can endlessly tread water by moving his Rook back and forth along that rank, White must eventually push his e-pawn.

3.e6

Threatening to win by 4.Kd6.

3…Rg1! and the coming cascade of Rook-checks will make it clear that White has no chance whatsoever of winning.

However, it occurred to me that White is not obligated to play 2.Ra7+. Instead, he can immediately push the e-pawn with check! The note I scribbled in my book years ago says: “This is not quite a Philidor position. How does Black answer 2.e6+?” I could not find a way to force either a Lucena position or a Philidor position, and regarded the position as simply “unclear”. So I turned the position over to my computer (Note: my computer is an ancient Fidelity “Master 2325” model from the 1990s, and it is probably not especially strong in the endgame compared to current models – but it’s the only chess computer I have), and let it play against itself after 1…Rg6 2.e6+!, with the following result:

2…Kc7 3.Ra7+ Kb6 4.e7 Rg8 5.Rd7 Re8 6.Ke6 Rh8 7.Rd8 Rh6+ 8.Kf5 Rh5+ 9.Kg4 and White wins. So what do you all think? Did Silman overlook 2.e6+, or am I (and my computer) missing something?

[Hal_Terrie]

@dennisk

I have found Jeremy Silman’s Complete Endgame Course: from Beginner to Master to be the most useful endgame book for me. I regularly go back to this book to review the basic endings so that I don’t forget how to play them. However, a few years back, I discovered what I believe to be a mistake in one of the problem solutions. But before I (a humble C player) accuse IM and respected author Silman of an error, I thought I would get some second opinions.

For those who have the book, it’s Problem 9 in Part Four (Endgames for Class “C”). Silman’s solution is given on page 153. I don’t know how to embed a diagram, but it’s a simple position: White has King on d5, Rook on a2, and Pawn on e5; Black has King on d7 and Rook on g8. It’s Black to move. Silman’s statement of the problem: “White threatens the unpleasant Ra7+. How can Black defend himself?” His solution follows:

If you thought you could draw by trading Rooks, you’re wrong: 1…Rc8 2.Ra7+ Rc7?? 3.Rxc7+ Kxc7 4.Ke6! followed by Kf7 is an easy win for White. Instead, Black can effortlessly hold a half point by using the classic Philidor formula.

1…Rg6! 2.Ra7+ Ke8

Now White’s King can’t advance onto the 6th rank. Since Black can endlessly tread water by moving his Rook back and forth along that rank, White must eventually push his e-pawn.

3.e6

Threatening to win by 4.Kd6.

3…Rg1! and the coming cascade of Rook-checks will make it clear that White has no chance whatsoever of winning.

However, it occurred to me that White is not obligated to play 2.Ra7+. Instead, he can immediately push the e-pawn with check! The note I scribbled in my book years ago says: “This is not quite a Philidor position. How does Black answer 2.e6+?” I could not find a way to force either a Lucena position or a Philidor position, and regarded the position as simply “unclear”. So I turned the position over to my computer (Note: my computer is an ancient Fidelity “Master 2325” model from the 1990s, and it is probably not especially strong in the endgame compared to current models – but it’s the only chess computer I have), and let it play against itself after 1…Rg6 2.e6+!, with the following result:

2…Kc7 3.Ra7+ Kb6 4.e7 Rg8 5.Rd7 Re8 6.Ke6 Rh8 7.Rd8 Rh6+ 8.Kf5 Rh5+ 9.Kg4 and White wins. So what do you all think? Did Silman overlook 2.e6+, or am I (and my computer) missing something?

You are (sort of) correct. This was probably a careless error by the author. The initial position can easily be corrected by placing the BK on e7 instead of d7, so that the pawn advance is not with check. Then the intended solution 1…Rg6 works.

The interesting part is that the actual position (with the BK on d7) is still a draw if Black plays the right move. Here is a hint: Black needs a move that will lead to regaining the tempo he knows he will lose because of the pawn move coming with check. See if you can find it.

– Hal Terrie

[mfschulte]

@Hal Terrie

The interesting part is that the actual position (with the BK on d7) is still a draw if Black plays the right move. Here is a hint: Black needs a move that will lead to regaining the tempo he knows he will lose because of the pawn move coming with check. See if you can find it.

It seems like there are actually a couple different approaches black can take here in order to draw – one easier than the other (at least to me).

[dennisk]

@Hal Terrie

The interesting part is that the actual position (with the BK on d7) is still a draw if Black plays the right move. Here is a hint: Black needs a move that will lead to regaining the tempo he knows he will lose because of the pawn move coming with check. See if you can find it.

– Hal Terrie

1…Rd8 seems to regain the tempo after 2.e6+ Ke7+. However, White is not obligated to play 2.e6+ either, and after 2.Ra7+ Ke8+ 3.Ke6, the White King succeeds in getting in front of his pawn. I’m not sure Black has a move that stops both White threats. I will admit, though, that (a) I’m not a strong player, and (b) I didn’t spend much time on this (because I don’t have much time).

[relyea]

FWIW, this is a TableBase position, so all modern engines know the solution(s) for all possible moves instantly. If you’d like to try ideas, https://lichess.org/analysis/standard/6r1/3k4/8/3KP3/8/8/R7/8_b_-_-#1

Alex Relyea