Lost in translation

I stumbled upon a slightly interesting endgame position analyzing one of my games from the USATS. The position in question never even came close to appearing in the game, and the position I'm presenting below has been modified further for clarity's sake. (I'll be using the Nunn Convention for my analytical symbols.)

Analysis position

White to move & draw

The important line runs 1 Ke4! Obviously White's king has to get within the square of the g-pawn. Kf7 The clearest line. 2 Kd3! White remains within the square of the g-pawn. g3 3 Ke2! e4 Or 3...g2 4 Kf2! e4 5 Kxg2 Kf6!= 4 Kf1! 4 Ke1? throws the game away. Kf6 5 Kg2! Kxf5 6 Kxg3!=.

What makes this slightly interesting (but just slightly) is that, as in so many cases, a slight shift changes the assessment completely. If all the pieces were moved down one rank, Black wins easily.*

Analysis position

White to move now loses

Here 1 Ke3 Kf6 2 Kd2 takes the White king out of g-pawns square.

Endgame books are littered with an uncounted number of examples where a shift by a file or rank makes all the difference in the world. And these things really do come up in our own games fairly frequently.** Being able to spot such differences several moves ahead in your analysis, and assess them correctly, will earn useful half-points over time. 

* In the original analysis position, moving everything one rank up would have given White good winning chances in a Q+pawn endgame.

** Except for rook & pawn versus rook endgames. For all the emphasis placed upon them, those pretty much never come up, at least in my experience!