Is the rook endgame really drawn?

by Sagar Shah - 10/08/2017

In the fourth round of the Chaleroi Open 2017, Sagar Shah played a rook endgame against WIM Andreea-Cristiana Navrotescu. In a rook endgame where she was two pawns down, Andreea resigned. Later when Sagar went back home and analyzed the position, he realized that the position was not so simple. He discussed the position with Tigran Gharamian, who couldn't find a win, and later with GM Sandipan Chanda. The position is still very rich and open for discussion. What do you think? Is it still a draw?

This position (given below) occurred in the fourth round of the Charleroi Open 2017 in Belgium. My opponent was the talented French WIM Andreea-Cristiana Navrotescu, who made her maiden WGM norm in the same tournament.

Andreea-Cristiana Navrotescu vs Sagar Shah

I took the pawn on f4 with my rook and my opponent resigned.

When my opponent stretched her hand in resignation, I did feel that it was slightly premature. However, I also had the belief that Black should win this position. Two pawns up in a very standard situation with no real threats should be good enough to win, right?

Well, I went back home and analyzed the position and to my great disbelief was unable to win the position. After the tournament ended I analyzed the endgame with my good friend and Levon Aronian's second Tigran Gharamian (2616). Tigran couldn't believe when I said that the position was drawn. He tried for next hour to break down White's defences but without any success. I present our analysis below.

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Analysis of the rook endgame"]
[Black "?"]
[Result "*"]
[Annotator "Sagar Shah"]
[SetUp "1"]
[FEN "8/8/6p1/5p1p/4kP1P/5r2/2R3K1/8 b - - 0 64"]
[PlyCount "22"]

64... Rxf4 (64... Rxf4 {And here my opponent resigned. And it seemed like the
right move. After all, she was two pawns down and Black had everything in
order. Surprisingly, and much to my disbelief this position turns out to be a
draw.}) 65. Ra2 $1 {The precise move to hold the draw. I cannot} (65. Kg3 $2 {
would lose the game in instructive way.} Rg4+ $1 (65... Rf3+ $2 {A bad move
that allows White to gain back the draw.} 66. Kg2 Ra3 67. Rc6 $1 Ra2+ 68. Kh3
f4 69. Rxg6 Kf3 (69... f3 70. Re6+ Kf4 71. Rf6+ Ke3 72. Re6+ Kf2 73. Rb6 $1 (
73. Kh2 $2 Kf1+ 74. Kg3 f2 $19) 73... Ra1 74. Rb2+ Kg1 75. Kg3 $11) 70. Rb6 {
also draws.} (70. Rg3+ $1 $11)) 66. Kh3 f4 $1 {[%cal Gg4g3,Ge4f5] is the other
way to win and a logical one. Black will play Rg3+ and get his king to g4.
This works because the rook is on c2 and the checking distance is not so huge.
If the rook were on a2, this would not work.} (66... Kd3 $2 {like in the game
throws the win away as after} 67. Ra2 f4 68. Ra3+ Kc2 69. Ra2+ Kb3 70. Ra6 Rg3+
71. Kh2 Kc3 72. Rd6 $1 {This very important motif of cutting off the king
gives White the draw.} Kc4 73. Rd8 Rd3 74. Rg8 Rd6 75. Rf8 $11) (66... Rg1 {
surprisingly wins, it's idea is to play f4 and Rh1+ and win the h4 pawn.} 67.
Rc4+ (67. Ra2 f4 68. Ra4+ Ke3 69. Ra3+ Kd2 70. Ra2+ Kc3 71. Ra4 Rh1+ {and the
h4 pawn is lost.}) (67. Kh2 Rd1 68. Rc6 f4 69. Rxg6 f3 70. Re6+ Kf4 71. Rf6+
Ke3 72. Re6+ Kf2 73. Ra6 Kf1 74. Kg3 f2 $19) 67... Ke3 68. Ra4 f4 69. Ra3+ Kd2
70. Ra2+ Kc3 71. Ra3+ Kb4 72. Ra6 Rh1+ $1 73. Kg2 Rxh4 74. Rxg6 Rg4+ $19) 67.
Rc4+ Kf3 (67... Ke3 68. Rc3+ Kd4 69. Rc2 Rg3+ 70. Kh2 Ke4 {[%cal Ge4f5,Gf5g4]})
68. Rc3+ Ke2 69. Rc2+ Kd3 70. Ra2 Rg3+ 71. Kh2 Ke4 $19 {The king comes back in
time and makes his way to g4.}) 65... Rg4+ 66. Kh3 f4 {The best attempt to try
and win, but doesn't work.} (66... Rg1 67. Ra4+ Ke5 68. Ra5+ Kd4 69. Kh2 $1 $11
(69. Ra4+ $2 Kc3 70. Ra3+ (70. Kh2 Rg4 $19) 70... Kb4 71. Ra8 Rh1+ $1 {A key
move.} 72. Kg3 f4+ $1 73. Kxf4 Rxh4+ $19 {And this is easily winning.})) 67.
Ra4+ Kd3 68. Ra3+ Kc2 69. Ra2+ Kb3 70. Ra6 Rg3+ 71. Kh2 Kc3 {This is the best
that Black can achieve.} 72. Rd6 $1 {I like this way of blocking the king.
Black is two pawns up, but has no way to make progress.} (72. Ra4 $11 {also
seems to draw.}) 72... Kc4 73. Rd8 Rd3 74. Rg8 Rd6 75. Rf8 $11 {Now these are
my findings of this rook endgame. I would really be interested if someone has
something more to add to it and is able to find a win for Black. As far as I
could analyze and also work with Tigran Gharamian after the tournament, we
came to the conclusion that the game was drawn.} *

The tournament was over, Tigran had won the event, but a dedicated chess player never misses out on an opportunity to learn

After the tournament in Belgium I travelled to Warsaw for my next event in Suwalki, Poland. I was going to stay one night in Warsaw and catch the train to Suwalki on the next day. As I reached the capital of Poland I received a mail from GM Sandipan Chanda, "Hope you have reached Warsaw safely. If you are not too tired, let's meet and recheck your endgame!" Sandipan was also in the same city!

Now I had travelled for nearly ten hours from Brussels to Warsaw, and was quite tired. But it was a great opportunity to see some new ideas in the position with a great analytical expert and of course I didn't miss it.

Sandipan met us with a smile and a pocket chess set!

After walking through the historical city for a few minutes and catching up with each other's life, Sandipan and I sat in Pizza Hut ordered an unlimited salad each and started our analysis!

Sandipan had read my report on ChessBase.com and had analyzed the position for some time on his own without using an engine. He shared his findings with me. (And yes, the salad wasn't enough!)

Sandipan's main idea was to first take the black king back to h6, then get the rook from e4-e8-h8-h7-e7 and free the king. And then slowly improve the position and get the following setup:

Black can get this position for sure. And it seems as if he is winning, because he can go Rd4 or even Ke5, threatening f4+. If Black gets f4+ and Kf5, he is winning. Sandipan's idea was not at all easy to refute.

However, later the correct defence was found. White should keep his king on f4 and rook on a4. Rd3 followed by Rh3 looked like a strong idea. However, White is fine after Ra6+ Kg7 Ra7+ Kh6 and Ra4! The point is to meet Rh3 with Ke5-Kf6 with a draw!

We tried many ideas during our dinner at Pizza Hut, but Black's defences are holding on. I would like to invite you to this discussion to put forth your ideas. What do you think? Is this a win for Black? The thing that I find extremely weird is that if this position is a theoretical draw, why hasn't it been covered in any popular endgame manual. Is it really a draw? Or am I missing something? I look forward to hearing from you in the comments section below.

A huge thanks to my two great friends Tigran Gharamian and Sandipan Chanda for spending time on this endgame and enriching my understanding.