The following rook endgame, and the way the players handled it, is very interesting. First it shows that even the best players often fail to grasp the subtleties of rook endgames; second, it is a good example with a very important idea to have in mind.
I must say that I first saw this position discussed in GM Alexander Khalifman's Facebook feed, so credit goes to him for bringing this to my attention.
The position below appeared in the game Aronian-Morozevich, from the World Championship tournament in Mexico in 2007.
It is Black to move and as usual in such cases it pays to think what White's plan can be.
The difference with the typical positions with an outside passed pawn is the e4-pawn - it is isolated and can be attacked and often taken by the white king. The fact that the pawn is on e4 gives White an important idea to play with - push the pawn to a7 (with a rook on a8), then maneuver with the king to win the pawn on e4 and then create a passed pawn on the f-file which will win the game by simply pushing it to f6. When the pawn arrives there the black king will either take it, allowing Rf8+ and a8Q or will move to f7, allowing Rh8 and after ...Rxa7, Rh7+, winning the rook on a7.
Understanding all this it becomes apparent that Black should stop the white king from wandering around.
Therefore, 35...Ra1! 36.Kg2 h5! would have solved the problem. Here the weak pawn on e4 has the important function of controlling the f3-square.
However, Morozevich played the immediate 35...h5? allowing White to activate the king with 36.Kf1.
Alas, not only Aronian didn't play that, he even ruined his chances of creation of a passed pawn on the f-file by playing the horrible 36.h4?. Now it doesn't even matter whether White takes the pawn on e4 - he cannot create a passed pawn on the f-file while a passed pawn on the other files are useless as they cannot force the black king away from the g7 or h7 squares. The game was drawn in 69 moves.
Rook endgames aren't easy. The reason is the need for both longterm planning and immediate calculation at the same time. These two factors are intermingled and influence each other and a failure in one aspect is often decisive. As we saw in the above example, even the best players often fail in this complex task.
P.S. From rook endgames to openings: the featured video was inspired by Gukesh’s choice against the Pirc Defence in the recently finished Norway Chess. He misplayed the opening, but I dig a bit deeper to understand what possibly he could have had in mind by choosing a line that has always been considered harmless for Black. It turned out that White can obtain a solid position and deprive Black of all the dynamism of the Pirc. Probably worth taking a look.