Nothing magical at all. Entering the position in Aquarium, Copy(pgn) in the ribbon, hitting pgn here, entering with ctrlC. Chessbase should have the same function.

Interesting engine evaluations by houdini 2 btw.

As I've said before, I know there's some pretty hot endgame analysts on this forum. This of course is a very famous knight endgame (I'm hoping someone can post the diagram I'm too stupid to learn how to do). I've never been able to bring myself to believe that Black's win was the inevitable thing many commentators (eg Keene, Muller) present it as, and I think my computer may have shown me why not.

So, at move 55 White has Kd4, Nh3, Ps e4 a3 b2. Black has Kd6 Nc6 Ps a4 b5 h4. Black obviously has just gone 55....Nc6+.

The game went 56 Ke3 Kc5 57 Kd3 b4 58 ab+ Kxb4 59 Kc2 Nd4+ 60 Kb1 Ne6 61 Ka2 Kc4 62 Ka3 Kd4 63 Kxa4 Kxe4 64 b4 Kf3 65 b5 Kg2 0-1.

I am interested instead in 56 Kc3. Muller gives now 56...Ke5 57 Kd3 Na5 58 Ng1 Kf4 59 e5 Nc6 60 e6 Kf5 61 Ke3 Kf6! (61..Kxe6 62 Nf3 etc) 62 Ke4 Nd8 63 Kd5 Nxe6 64 Kc6 Kf5 65 Kxb5 Kg4 66 Kxa4 Kg3, and Black wins.

My machine, however, wants to play 61 Kd2 (!!, I am tempted to add) After 61...Kf6 it plays 62 Kc3 (I fancy this may be a mutual zugzwang position; if White's king were on d2 then ...Ne5 would win handily). As it is, though, so far as I can see Black really can't do much here, but I wonder if any more perceptive reader can propose a way of overcoming this defence?