The difference of a square can cost you an endgame 14 moves later

I recently lost a winning endgame and had to check the engine analysis to understand where I went wrong. It turns out the engine claims there's only one obscure move that wins. This was the position:

In this position, the engine claims the only move is Rc3! I got intrigued and wondered why a different rook move on the third row wouldn't do the trick. Rb3 looked all the same to me. My curiosity led me to a 14-move sequence in the Rc3 line and almost an exact 14-move sequence in the Rb3 line.

Rc3 line:

All the moves for black are the best according to the engine, except Kh3 at move 11 which is just a faster mate for white. Interestingly the same position (only with the white rook on b3), Kh3 keeps the position in a draw while Kh4 (the better move for black in the Rc3 position) will lead to a win for white. Nevertheless, I chose the Kh3 move to show how the difference of a square for the rook can completely change the game's outcome.

Let's see the Rb3 line:

When playing the same moves in this position I was perplexed as the engine kept showing a draw even though it looked almost identical to the Rc3 position. Only 13 moves in I realized the difference: the rook wouldn't be protected if it was moved to the first row!

What's the moral of the story? Sometimes the engine's analysis is not as simple as it may seem and even if the position is at +5, it might as well be a draw.

Ok

That's bad for you

Well yeah thats chess engines for you.

A human would get to the same conclusion as the engine, only it would take much more time.

A few years ago I was studying a position from a book (but not a game of mine) along with three other players.

I told the three of them: "What's the difference? They are just one file apart, they are all the same." (I was referencing this to a king's ability to attack a pawn in a middlegame)

Many moves later, I finally realised that the file made a huge difference in an opposing king (White) reaching a promoting pawn (Black), in the sense that the file was necessary to keep the White king from reaching the promoting square in time. I then remarked "Why does Black win? Because they are one file apart", and the three other players laughed.

What a silly remark by me, ending in me making the same comment of being "one file apart" after realising the difference.

I am no GM and thus I dont know if a GM could do this with such extreme precision.