What is clearly working
You have a real identity as a player. The flank and kingside fianchetto systems you live in, the King's Indian Attack, your Modern setups and your Czech Defense results, are not just comfortable for you, they clearly fit how you like to play. You build a solid shell, then you switch into attack mode once your opponent gives you a target.
Your win over Paul Sanchez is the cleanest example of that style. You castled long, pushed your h and g pawns at his king, and finished with a queen checkmate on the seventh rank. That is a mature attacking game, not a random pawn storm. You can review it here: your checkmate win.
The win against LuckyMajky shows the other half of your game. You were happy to accept an exchange, then you slowly improved your rook and king, pushed your outside pawn, and won a long game without needing any fireworks. That patience is a big asset.
Pattern one: rook endgames are where points leak
Both of your recent losses ended in a rook endgame where you were the one under pressure.
Against nissou-ach you built a nice clamp on the kingside with your pawns marching up the board, but once the pieces came off you were a pawn down in a rook ending and never got that pawn back. Against Rodwell Makoto the game went deep into a rook and pawn battle where every small decision mattered and you eventually had to resign.
The habit to change is this: in rook endings, do not count material first. Count activity first. An active rook and an active king are often worth a pawn, and sometimes more. Before you grab a pawn, ask whether your rook will still have a square to attack from afterwards. If the answer is no, the pawn is usually not worth taking. Review both games with that single question in mind: this rook endgame loss and this long endgame.
Pattern two: the clock decides too many blitz games
In several of these games your clock dropped into the last twenty or thirty seconds while your opponent still had a comfortable minute. In a three minute game, that gap is the whole game.
You know your openings extremely well by now, so use that. Play the first eight to ten moves quickly on autopilot and bank that time deliberately. Then spend it where it actually matters: two or three critical moments in the middlegame where you have to make a real choice. You do not need to think on every move, you need to think on the right moves. Treating time trouble as a decision-making problem rather than a speed problem usually fixes it fastest.
Pattern three: king safety before material
In your draw against Gabriel Flom you were winning material, you took the knight, and yet the game still ended in a draw because his queen had a permanent checking pattern against your king. That is a perpetual check you want to see one move earlier.
When you are ahead in material, the question flips. It is no longer "can I win more?" It is "can my opponent check me forever?" If the answer is even close to yes, spend one tempo on king safety, give your king an escape square, or offer the queen trade. A single quiet move there converts a draw into a full point. You can see how the game ended here: the draw by repetition.
Repertoire notes
Your results are strongest when you stay inside a small group of structures. The Czech Defence and the King's Indian Attack clusters are doing the heavy lifting, and your scoring there is excellent. The gambit lines you play a lot, like the Amar Gambit and the Hungarian Wiedenhagen-Beta Gambit, are doing fine too, but they are a large slice of your games.
My suggestion is to keep those gambits as surprise weapons rather than a default plan. Pick two main systems that you already trust, and keep one gambit for opponents who clearly want a slow, quiet game. Also keep working on prophylaxis in your Modern setups: when your opponent refuses to commit in the centre, the winning plan is usually to find the central break yourself rather than to keep improving pieces forever.
What to work on next
- Spend a week on rook endgames only. Learn the Lucena Position and the Philidor drawing method until you can play them from memory, then apply the activity-before-material rule in your blitz games.
- Adopt a clock habit: fast opening, slow critical moments. Aim to reach move fifteen with more time than your opponent, not less.
- Before every material grab when you are ahead, scan for checks against your king. If there are two or more, deal with them first.
- Keep playing the attacking games, they are your best chess. The Paul Sanchez win and the LuckyMajky conversion are your template: build, attack, then convert calmly.
- Do a short review after each session and write down one endgame position you got wrong. Ten minutes of review per session adds up fast at your level.
You are clearly closer to your peak than your recent dip suggests. The attacking play is already there, the missing points are sitting in rook endings and in the last thirty seconds on the clock.