ambiguous conclusion, can anyone help?

Hi, I was studying one of my endgame positions (using online shredder endgame database).

It seems there is an ambiguous result for one position (see 2 images below). The only difference is the board is reversed (a1 square is at upper right in one, and lower left in other).  Both cases have "white to move".  In first conclusion, the game is a draw.  In the other white loses in 19.  Is there anything I'm missing in reading these results?

Thanks if anyone can help.

btw, I hope it's OK to ask a question about shredder here (a separate website).  I could not find a forum there, so I came here to ask.  Thanks.

You're not missing anything. In your first example, white is going to promote so black has to give up the rook for the pawn.

In the other, white won't be able to promote, apparently, likely due to the fact black's king is able to get into the action quickly enough to force the win.

Pawns only move 1 direction.

Thanks ripachu and martin for helping me solve the mystery!!  It seems when I went to shredder, I should have first switched the board (since I was playing from the black side) before entering the piece positions.  (I was mostly curious to see if I know how to read the endgame tablebase, rather than evaluating my game - but in my blunder just ended up confusing myself!!!)  Well - now I know how to use that site!

If anyone is curious to see the game where I came to this position it is here (with some comments)

Thanks again!

"Here, 30 moves later, white gets back the gambit pawn"

Move 30 was roughly the 10th time (!) white had the opportunity to win back the pawn that game.

Yes I noticed that too.  Weak pawns are often weak for a long time since they can't change files easily, and can't go backwards (like you mentioned).  I suppose white just took it when it was convenient (or when we both felt it was advantageous to trade other material). Thanks for your observation

I thought maybe you didn't notice it until then, never mind then