Accuracy of the evaluation bar!!??

Today, I played one of my first long chess games on chess.com (30 minutes). This is rare as I mostly play daily, bullet and blitz on this website. Don't get me wrong, I do play long chess games but most of it is over the board or on another site that I use.

Whenever I play long chess games, I always analyse it to the max. Quite literally... with another website, chess.com and Stockfish 13. However, I am starting to doubt the accuracy of the evaluation bar after today's game.

Before I explain why, let me show you the game (it can also be found on my profile, already analysed).

At move 13, after I make my move it says the position is basically draw. An evaluation score close to 0. However, I cannot fathom the idea of the game being a draw when I am a whole 4 points ahead (1 bishop and 1 pawn). How can this possibly be a draw? When I'm watching commentators analysing games, they talk about how being 1 pawn down is huge for grandmasters. 1 bishop is unwinnable.

So how is it possible for a game to be drawn when one player has a massive material advantage. I know computers are meant to be smart, a lot smarter than humans but should I continue to put my trust in computers because it doesn't seem like its very accurate.

Hmm, black has an attack against your king. The computers are telling you that the severity of the attack offsets your material advantage.

That should not be hard to understand. Remove your knight on e5 and you would get mated in two moves.

Yes but the evaluation bar calculates the scenario where both players play the best moves for the rest of the game. This would assume that I don't do anything stupid like moving my knight from e5, right?

The exposed king is an exploitable feature for black, and at depth, the position is winning for black, not even. The knight at e5 is immediately attacked.

I think the H pawn has a role in protecting white from checkmate. On move, I think, 16, the G pawn moved to g3 to block check, right, and the only defender of the pawn was the H pawn on h2. If that were to be removed, it would also result in a mate in two.

Essentially there is much more to a position than material, which becomes clear when you get to the higher levels. Factors such as initiative, space advantage, pawn play, etc... are extremely important. A similar principle is also evident in King v King and Bishop, even though one player has a +3 material advanage, it is impossible to win or lose. To address the point about GM games, the positions are usually winning when one player is up just a pawn because the other (previously mentioned) factors don't outweigh the material, and the skill of super GM's is so high that they are able to convert very small advantages, even if the engine says +0. Long story short, the engine could beat Magnus Carlsen 10000/10000 at a relatively low depth.