I added some analysis with 53...b4 which is not all inclusive, but I did demonstrate one win and one draw. In the meantime, I'm not taking Fritz's word for 53...b4 winning in all lines until I see them myself, I have F12 and will see what it plays. Still, all these sidelines need examining, after awhile I started to lose track of them during the game!

53. ...b4 (instead of c4) wins according to fritz. reason is that after 54. c4 white has no more tempo moves (black has two left), so black can first push the white king back and then after playing at one point ...a4 ba Kxc4, the a4 pawn eventually falls

61.ab, black wins with something like 61.ab a5 62.ba Ka5 63.Ka3 Kb6 64.Kb4 Kc6 65.Ka3 Kc5 66.Kb2 Kd5 67.Kc2 Ke4 68.Kd2 Kf3 69.Kd1 Ke3 where black is just going to win the c-pawn. mattgross

64.Kb2 maintaing the distant opposition also maintains the draw IMHO

After the knight trade, this is just a trivial win for black, right?  It looks to me that black went wrong here by not liquidating the doubled pawn when he had the chance.  The last point where I see a win is move 60, where black played 60... a5.  If instead black simply played 60... cb, I think the win is pretty easy.  On 61. cb, black just runs the king back around (61.cb Kc6 62. Kc2 Kd5 63. Kd1 Kd4 64. Kd2 c3 65.Kc1 Kd3 66.Kd1 c2 67.Kc1 Kc3 68.a4 ba 69.b5 a3 ba 70.a2 a7 71.a1(Q)++ is a sample line), while on 61.ab, black wins with something like 61.ab a5 62.ba Ka5 63.Ka3 Kb6 64.Kb4 Kc6 65.Ka3 Kc5 66.Kb2 Kd5 67.Kc2 Ke4 68.Kd2 Kf3 69.Kd1 Ke3 where black is just going to win the c-pawn.

Pawn-up king endings are usually won for the side with the extra pawn, as long as it isn't one pawn vs. none.  Black's sin here was getting to a position where he needed to trade all the pawns off in order to make progress.  Doubled pawns can be a problem when you can't liquidate them, but even though he could have done so easily - he just didn't.

considering chess.com computer is only 2500 strength and a good engine is about 3000-3300 (theoretically) and especially when you take into account how engines typically excell in tactical realm its no wonder this computer is so mediocre in the endgame. 

I think the chess.com PC is mediocre at best at least for endgame analysis for instance it's evaluations here in the K+P endgame stop at several instances when Black has the distant opposition which the PC rates as -+ regardless of the fact that the distant opposition is seemingly irrelevant.

I'll eventually get Fritz12 to evaluate this game and post that as well

However, the position after the Knights trade is very interesting ! Does the extra doubled pawn win ? Or at least, how to make white's life as miserable as possible ?

The chess.com computer is infamously bad at judging anything more than quick tactical equasions.  

I couldn't find the win either but I'm not very good at endgames so.. :/

Is this me or is chess.com computer completely off-base assessing the pawn endgame after 61...Kxa5 as better for black ? I can't see how black is supposed to win after the computer's recommended 62...Ka6