Saturday, April 14, 2018

Endgame Analysis: Rook Endgames With an Extra Pawn

Hello again everyone, and welcome to my latest article on the endgame. It's probably been well over three months since the last time I published an endgame article.

Just to give perspective on how this article has come about, has anybody ever told you that chess is a lot like the news? Take The Rachel Maddow Show (Weeknights, 9pm Eastern on MSNBC). Many times, she has started her show with a statement that she had many topics planned for her show that night, but one or two breaking news items in the last 30 minutes has just completely wiped all that out, and the next hour is spent covering those breaking news items.

Well, in many ways, that's what happened here. Originally, I was going to publish the fourth edition of The French Connection, but that is going to have to wait until the next publication. Instead, we are going to be looking at a rook endgame that happened this past Tuesday night. The tricky thing about rook endgames is that even the slightest of errors can completely change the assessment of the position. We are going to look at one here where a couple of wrong moves were made by both sides, but in the end it was White, who was probably "better but not winning" going into the endgame being up the pawn, won.

With that said, let's see what happened.

Tuesday Night Action 41, Round 5
W: Patrick McCartney (2049)
B: Adityz Shivapooja (1887)

We are going to start at the position where the Rook endgame virtually began.

Position after 32...Ra4

First let's make a few assessments about the position:
  • Black will grab a pawn. There is no way for White to save both the a-pawn and the c-pawn as 33.Rb2 will be answered by 33...Rc4!, winning the c-pawn.
  • The only way for White to get a pawn while Black is getting his is to go for the c-pawn. This is going to result in a case of White having his Rook in front of the passed Pawn. Before White will ever be able to get behind it, Black will reach that point first by going to the c-file. With his own passed a-pawn to boot, these pawns are eventually going to get traded.
  • Rook and 4 Pawns versus Rook and 3 Pawns, namely e-, f-, g-, and h- versus f-, g-, and h-, is often viewed as a theoretical draw. However, here we have Rook and 3 Pawns versus Rook and 2 Pawns with Black's pawns being split on the kingside.
So let's absorb what we have here. White doesn't want to outright lose the extra pawn and give only Black a passed pawn. Therefore, he will have to go for the c-pawn. This is follow up with a trade of queenside pawns, and so therefore, White has to focus his attention on the kingside majority if he wants any chance at winning. This pretty much drives the idea of what happened the next 10 moves or so.

33.Rc7 Rxa2 34.Rxc6+ f6

Often times, if you are the side defending in a pawn down Rook endgame, you want to keep your most centralized pawn on the 7th rank if possible. It can act as a shield to your King on the seventh rank (second if you are White), and with the King on the seventh, the opposing Rook cannot check you from behind as you will just capture it. With this advancement of the f-pawn, Black always has to look out for the White Rook invading via the back rank, and in some ways, this is the root cause of what ultimately forces the Black King himself to get stuck on the back rank. I personally would play 34...Kg7, but the move played may not lose because it sets up a fortress that White will have a hard time breaking if Black uses the Rook purely to guard the seventh rank and simply toggles until White tries to do something as we will see a few moves down the road.

35.Kg2 Ra3 36.g4

Advancing the central pawn first, avoiding situations like 36.h4 h5 or 36.f4 f5.

36...Rc3 37.h4 a5 38.f4 Kg7

There was no immediate mate threat, but Black did have to watch out as White was about to play 39.Rc7, at which point, Black would be forced to further advance one of his pawns to avoid checkmate. or loss of the Rook. That said, 38...Kf7 might be a slight improvement, getting out of there and centralizing the King, intending to answer 39.Rc7+ with 39...Ke6. If White goes quickly running after the h-pawn, Black grabs the c-pawn and still has his own passed a-pawn on the side of the board away from the Kings.


Now instead, Black is forced to the back rank.

39...Kg8 40.c6

White wants to keep his advantage as mobile as possible. Advancing the h-pawn leads to difficulties making progress. For example, 40.h5 Kf8 41.c6 a4 42.Kf2 a3 43.Ra7 Ke8 44.Ra8+ Kf7 45.c7 Rxc7 46.Rxa3 Rb7 47.Kf3 Re7 48.Ra1 Ke6 49.Ra6+ Kf7 50.Kf2 Rb7 51.Kg2 Rb3 52.Ra7+ Kg8 53.Ra8+ Kf7 54.Ra7+ Kg8 55.Ra8+ Kf7 and it's very hard to see how White makes progress.

40...a4 41.Kf2 a3 42.Ra7

Note the timing of White going to the a-file. This is very important. With White's pawns advanced, we saw in the note to 40.h5 that one of White's main problems was the cutting off of the King by Black on the third rank. If White played Ra7 a move earlier, then Black could grab the c-pawn, and once White grabs the a-pawn, go back to the third rank to cut off the White King. Here, Black can try to check the White King and then push a2, guarding laterally, but eventually will have to take the White c-pawn anyway, or he can trade immediately like he does in the game, but the specific location of the White Rook following the trade allows White to lift the King past the second rank.

42...Rxc6 43.Rxa3 Rc4 44.Ke3

In many ways, White is trying to make something out of nothing, but when you are a pawn up and have nothing to lose, being a pest is sometimes the best way to go. Test Black on his defense and make him prove that he sees that there is nothing that White can do.

44...Kf7 45.Ra7+ Kg8 46.Rd7 Ra4 47.Rd4 Ra7 48.Ke4 Kf7 49.Rd5

So White has been able to bully his way to the fifth rank, but with correct defense, he should get no further.


Black correctly continues to toggle on the seventh rank, daring White to try to prove he has something.


White's ultimate goal is to get the King to d6, which he cannot do with the Rook on d5, so he shifts the Rook over a square to stay in close proximity with the King. 50.Kf5 leads to nothing. Black can toggle on the seventh and White can toggle on the fifth all they want. The game goes nowhere.


This move in and of itself doesn't lose, but why mess with it? Better is 50...Ke6, immediately asking White what he's going to do. After 51.Rc6+ Kf7, White has nothing better than to go back with 52.Rc5. If White tries to charge forward with 52.Kd5, then 52...Rb5+ and if 53.Kd6, then 53...h5! puts an end to all hopes of White winning. The White h-pawn falls. If 54.g5, then 54...fxg5 55.fxg5 Rb4 wins the h-pawn, and any trade on h5 leads to an easy winning of the h-pawn and a drawn position.

51.Kf5 Rb7

Correctly returning to where it needs to go as long as the White King stays in proximity with the pawns.


Playing 52.h5 first does White no favors. Sure, it gives the King the h4 square as a hiding point without dropping any pawns if Black goes on continuously checking White once he moves the Rook to c6, but what is the King going to do on h4 anyway?

A word of note to those that use computers. Many computers have been thinking that White has a winning advantage the whole time, and some see this move as the blunder that allows Black to equalize. You can't trust computers in openings and endgames, and in reality, Black has always had the draw with correct play, it's just that now computers realize it.

52... Rb5+ 53.Ke4 Rb4+ 54.Ke3 Rb3+ 55.Kd4 Rg3 56.Rc7+ Kg8 57.Rc8+ Kg7 58.Rc7+ Kg8


White makes one last ditch effort to win the game. Remember what I said about being a pest when you've got nothing to lose? Well, guess what, this one worked! Black finally buckled!

59...Rxg4 60.h5

The trick is that this pawn is forever tactically defended and hence poisonous at all times unless White retreats the King back to the fourth rank.

60...Rh4 61.Kf5

Black to Move and Draw


This move loses now and will always lose! Black must play 61...Kf8! and White can't make progress. For example, if he waits and plays something like 62.Ra7, Black just continues to toggle between f8 and g8 with his King. Any check by White? Black simply toggles between the seventh and eighth rank, specifically on the f- or g-file, so as not to allow the Rook to ever capture the h-pawn. The moment that the White King tries to charge, Black either checks or captures the f-pawn based on the following circumstances:
  • If the Black King is on g8 and the White King goes to g6, Black checks from g4.
  • If the Black King is on f8 and the White King goes to g6, Black takes on f4.
  • If the White King takes on f6, then no matter which square the Black King is on, Black will take the f-pawn with check.

Following this script, White can achieve nothing better than Rook and h-pawn versus Rook, which is a draw. Note that Black can NEVER take the h-pawn, even with the King on f8. For example, after 61...Kf8 62.Ra7 Rxh5+?? 63.Kxf6 Ke8 (63...Kg8 leads back to the same problem as what happens in the game after 64.Kg6) 64.Ke6 and now both 64...Kd8 f5 and 64...Kf8 65.Ra8+ Kg7 66.f5 Rh1 67.f6+ Kg6 68.Rg8+ Kh5 69.f7 Rf1 70.f8=Q Rxf8 71.Rxf8 Kg4 72.Ke5 h5 73.Ke4 are winning for White. In the latter case, no matter what Black does, 74.Rg8+ is coming next.

Instead, the move played in the game loses instantly.

62.Kg6 1-0

So as we saw in this game, Rook endgames are the one endgame where being a pawn up is frequently insufficient to win, but the side down the pawn must always make accurate moves, and so the player up the pawn usually has nothing to lose by playing on.

No comments:

Post a Comment