Most of my games go as follows: good opening, gain some advantage, lose it all by 1 bad move

How am I supposed to improve knowing this? I play a lot of tactics and try to improve my positional play but I feel like the middle game is so complex with so many directions that unlike openings I don't have a good grip on it no matter how many puzzles I do. Here is an example from my last game. Also note that I always have a time disadvantage because I think through early in the game giving me the advantage, but that also gives me time trouble after which I am forced to think quickly and throw.

First off, it looks like your time management was not too bad until moves 24 to 29. And I didn’t really feel like those moves were particularly complex, so I’m wondering why you took so much time to play those. In a 10 minute game, you don’t really have the luxury of spending 30 seconds for several moves in a row. If you want that kind of time, maybe you should play 30 minute games.

Middlegames are also extremely complex, so in general I don’t think anybody has a great handle on all of the things that are happening on the board at once. Probably the best we can do is try to find safe moves that put pressure on the opponent. To be clear, you didn’t lose this game because of one bad move, though. You lost this one because your minor pieces were never putting much pressure on black. You didn’t play Ng4+ when you had the chance, your LSB was just stuck for a lot of the game, white’s rooks got active and started pestering you everywhere.

There’s a few ways you can improve. If you specifically want to improve on time control, you could look through portions of a game like moves 23-29, see why you spent so much time on these moves, and then maybe do some training to find these moves faster. Maybe that’s just finding out how to identify and be satisfied with safe moves that put pressure on the opponent.

If you’re trying to hang your pieces less, something I’ve been finding useful is going through a grandmaster game, playing from the perspective of the player who wins, and asking myself before playing every move, “what would I play here, and am I hanging any pieces?” It’s easier to do this kind of practice when you have unlimited time to think for each move, and it’s less stressful because you won’t actually lose games if you play the bad moves only during practice.

If it’s a mindset thing, if you notice that you start to play carelessly once you’re ahead, I think the remedy is just to understand how advantages are usually converted. They usually take a long time, and often the side with the advantage is looking for ways to limit counterplay, find the most useful thing the opponent can do and make sure the opponent can’t do that anymore.

Just a few suggestions, hope they help.

"And I didn’t really feel like those moves were particularly complex, so I’m wondering why you took so much time to play those."

IDK why capturing the knight took me so long, but the move Dd4 that I thought about for a while turned out to be a blunder according to the engine. I saw Dd2 but after the king sidesteps I didn't see a particular threat, and I would lose sight of the b6 pawn.

"You didn’t play Ng4+ when you had the chance"

In hindsight it seemed more obvious but after the king would sidestep, perhaps Kg1, I didn't see a good continuation in time.

"If you specifically want to improve on time control, you could look through portions of a game like moves 23-29, see why you spent so much time on these moves, and then maybe do some training to find these moves faster."

What kind of training would allow me to practice finding such moves faster? I feel like they are heavily dependent on the position, and they don't appear in puzzles. I can analyse this game and try to follow my own thought process and correct for blind spots or missing tactics, but I am not sure if I would notice something similar again in another (similar) game I play a few months ahead.

"If it’s a mindset thing"

Well I do play more recklessly once I am at a clear disadvantage. Complicate and they will blunder kind of thing. Except I blunder.

Training to find these types of moves faster... that's been a struggle for me, too. And unfortunately, I don't have a great solution. But I do feel like the best way to work on time managment is knowing when a position is "critical." A critical position is when you need to find the best or the second-best move, and all other moves will make you worse. You know that a position is critical either when your position is starting to get worse, when your opponent is about to make a move that will seriously improve their position (free up their pieces), or when there are a lot of tactics on the board, and if you don't resolve them favorably, you will lose a lot of material. Probably the best way to learn when you need to find the perfect move is to look over your games and try to find those positions where your games started to turn sour. Not just due to tactical oversights (though those are important too), but also especially at those moments where it just started to become harder for you to move your pieces and easier for your opponent to move their pieces. If you're using an engine, often this is the point where the evaluation for your advantage starts to decrease, or where there are fewer and fewer moves that lead to an advantage for you.

Another way to find moves faster, though, is to play over grandmaster games. Grandmasters will play good moves that you or I typically won't find, so just seeing them and seeing that those maneuvers are possible should give you more of a chance of seeing those types of moves in your own games. The more GM games you play over, the more you'll be able to identify good moves you haven't really noticed before.

If you want to notice something similar in another similar game that you play a few months later, I think that's what I would do.

As for the mindset once you have an advantage -- what you had to say is very revealing. When you have the advantage, you should not complicate the position. Instead, you should recognize the nature of your advantage and then steer the position in a way that allows you to exploit that advantage. For example, early in this game, you had a material advantage, specifically, an extra piece. A good way to exploit an extra piece is to make sure that your extra piece is doing work. That's why the engine recommends 15... Bf8 or 15... g6 -- the engine recognizes that your DSB isn't doing any work, so it wants to find a way for that bishop to get into the game (via ...g6 and ...Be7-f8-g7).

@1
Your decisive error was 37...Kg6? while 37...Nc6 would have saved material and held the draw.
36 Rb3? lets you back into the game. 36 Rb2 was winning.
34...Kh7? loses. Why not take the pawn 34...Bxc4
33 bxa6? lets you back into the game. Taking the 7th rank with 33 Re7 was devastating.
31...Qb4? loses to tactics, you had to defend 31...Qd6
30...Nxd5? gives up a piece without reason. Just 30...axb5 was still winning.
27...Rb8? is a missed win. 27...Ba8 was necessary.
26...Qd4 is not wrong, but 26...Qd2+ was crushing. If you had spent more time here, you might have won right away.
11 Nxd4?? blunders a piece to a fork
5...d6? was weakening and loss of time. Bc8 is developed, now develop Bf8: 5...g6.

You had Nc6 for several moves and you didn't play it. That's why you lost.

Hi! It is frustrating to lose a game by a blunder when you should have won strategically instead. But chess is exactly like that - and probably this is one of the reasons people likes it the most. So you have to be alert the whole game and not be confused no matter the advantage you may have : material, time, position...Then you should always think in ways to lose, and avoid them. Not easy, happens at any level, but one can be trained to improve this.

Good luck!

Good try! Converting is a game itself that I am slowly improving on.

Also good to see a fellow Katalimov player. Best of luck.

It is not so easy to prevent that from happening, but always paying due attention to all the opponent's resources can reduce those losses.

I have to learn from this

Almost everyone's games go like that a lot of the time