Desperate for chess improvement

Hello, chess.com users! 

I've been drown in sorrow those last weeks. It seems like i am officially stuck again, unable to improve. My peak rating is 1972 - which i achieved about a month ago, i started outplaying my opponents and they seemed all weak. a little time after - all my rating is gone, i reached the miserable low 1800 again (~1820), playing super badly and commiting obvious mistakes. i played like this for a week and decided to take a one day break. it worked. i got back to high 1800, but it seems like i cannot improve past this. I'd love some advices on going through this dark era.

My main strenghts :

- 2100 Strength positional chess understanding

- Good at endgames (and general endgame strategy)

- Ok tactics and nice puzzle-solving skill

- Nice opening knowleadge (by far not my main focus)

My main weakness :

- Intuitive play. I do not think in a system way. My opponents often play moves that i didnt even realized were there and then i just play the position. What is your thought process?

I always play with frequency here : every - single - day and i always analyse my game afterwards it's done. It just feels like to get past 1900 - 2000 you need some unknown obscure knowleadge!

In your most recent game you lost:

https://www.chess.com/game/live/110876433077?

the main thing that I noticed:

In the opening, you played the move f5 which seemingly wins the g2 pawn and the rook on h1, which turned out to be not the case. Because you are playing 30 min, if you play a move like f5, you have to be willing to calculate the line all the way to the end, to figure out its conclusion, you can't just assume that since you are winning the g2 pawn you will win the rook and the game. Another thing i noticed with that was that you took 20 seconds to play that move, which means that you maybe saw it and calculated 1 or 2 moves.

An example line I might've calculated for this:

f5, ng5, bxg2, nxe6, qe7, qe2(threatening nxc7), bxh1, nxc7, kd7, nxa8, qxe2, kxe2, be4, saving my bishop.

The position then enters a dynamic state, as white is up a pawn but the knight on a8 is trapped. Just from intuition, I would assume that White has some way to save the knight as black needs 4 moves to save it(nf6,be7,nc6,rxa8), so I would assume that White is better.

If i could, I would calculate as far as I could, or until I saw a winning line for White that would invalidate the whole thing.

You should study openings.

Yes, i played f5. I'm not a crazy random dude though, i saw after a game i won i had the chance to play f5 and be up a full point in the opening according to the engine. it's weird. i guess c3 makes everything different. i could have played way safer but i couldnt forget that other miss in the other game (where f5 was winning and therefore was correct) .

Being stuck does not mean you aren't improving plus you opened your account about a year ago 1900 elo in about 12-24 months is decent I was stuck at 2000 for about 8 months until I started gaining elo part of it is due to all the 2000s being talented at chess the higher the elo usually the faster they learn chess which makes your opponents harder to beat cause you have to learn faster then them

I actually recommend trying to get better at a different time control other than rapid usually if your stuck at one time control you can improve in another if I'm stuck at bullet I switch to blitz or rapid stuck at blitz switch to rapid or bullet