I don't trust Computer's analysis of the game ending.

While I played the end of this game... I was very optimistic... I constantly felt that I could win.

Is this self-deception? Or I really could?

Your opinion?

I hope someone else can chime in. I really don't know.

Ok.

In fact I had only 3 minutes left in my disposal... So I couldn't even think over the position well.

I know. A 10|5 would be nice.

After 26. Kg2, it's pretty hard to imagine winning this endgame for either side. But your opponent went a bit wrong and gave you chances.  32.e6 was a very impulsive move, and I'm not sure what you were intending with that move. Endings with nothing but bishops of opposite colors are famously hard to win, even with an extra pawn

Won? You certainly could have lost after 42. ... a5! I see no reason to doubt the engine analysis at that stage. Even for a human it is quite obvious. I noticed that the engine gave you a good chance after 17. Bf4 but that was a long way back in the game.

Hmm... then it's strange.

I actually felt my advantage in the end of the game.

My intuition constantly was about that I could win. )

Funny.

GM Ginger explained that a player should trust his intuition.

I meant if I moved 39. Bishop b4.... instead of 39. g4... and only then I did send my g-pawn ahead for promotion... So my plan was.

My left flank was solid and in security then... and I could move quietly my g-pawn to turn into Queen.

After 34. Bc7 the game is drawn, but somehow you managed to get losing chances later. Fortunately for you, your opponent did not capitalize.

It is ok not to trust this quick computer analysis. They are quite shallow, and can actually be wrong. In these cases, use your own engine and give it more time.

... during play. After it, your intuition requires nourishment with a piece of solid analysis.

Ok. )

opposite color bishop endgames with equal pawns are drawn most of the time

though 42. a5! would have won the game

Yes ! So I could win though... )

I felt it. But... as always... when only 3 minutes remained on the clock... I begin to hurry.

He meant 42...a5! could win for black.  Sorry to disappoint.

Sometimes one can draw even two pawns down in an opposite colored bishops endgame.

Here is an example.

    doneevin   haver   cumpewtur    h   at da   liebry 

 y ya   sideweez ?

I do believe there is a tablebase option that isn't wrong for the ending, not sure if that applies here, was it 7 pieces total or for each side? hmm...

If there's only 7 pieces then, use tablebase: https://syzygy-tables.info/