t3gantry vs toan6ca

for ♞ David Claude 2026-07-18

The story of the game

David, this is the revenge arc: on Tuesday a knight on h5 lost you a game; tonight 22...Nh5 won you one — because this time it had a target. You wobbled a little in the middlegame (a pawn's worth after 10...g6), stabilized, and then finished with the exact move your review asked of you: the in-between blow. 23...Nxg3 didn't answer the attack on your queen — it asked a bigger question first, at 42 seconds on the clock.

The key moments

10...g6 — the one real dent (53% → 40%). The engine wanted 10...e4 — the pawn push with tempo against the f3-knight — over a quiet weakening move. The same theme flickers at 13...Qd7 (the engine liked the active Bg4) and at 18...e4, played one move after its best moment (18...a5 first, prying the queenside).

7...O-O — the bishop that forgot its job. 3...Bb4 exists to take on c3 at the right moment; the engine's choice here was exactly 7...Bxc3. After the retreat to a5, White got a3 and b4 in and the bishop watched the rest from b6.

22...Nh5 — the rim knight with a job. The difference from Tuesday: this one attacked something (the lifted rook) and had a follow-up. Rim knights aren't bad — aimless ones are.

23...Nxg3 — the zwischenzug. White's 23.Bg4 attacked your queen and assumed she'd move. You checked the forcing moves first and took the rook. That is, verbatim, the "counterplay before defense" drill from your profile — executed in under a minute.

Profile check

Log the pattern: three of your four flags were moments where an active pawn move was the engine's first choice. Your winning blow was a piece capture — but the games you'll simplify into wins come from finding ...e4 and ...a5 on time.

What to drill next

Every move, one question before placing a piece: "is there a pawn move with tempo?"
When you play ...Bb4 systems, taking on c3 is the plan — don't preserve the bishop into passivity.
Keep the zwischenzug scan: attacked ≠ obligated. You proved it tonight.

for ♝ James Claude 2026-07-18

The story of the game

James, the painful and the promising in one game: you drifted from +1 to level without a single tactic being played, were still dead equal at move 23 — 47.5% — and then lost everything to one in-between move at 49 seconds on the clock. The collapse window is down to single moves now. That's progress wearing a disguise.

The key moments

6.Nc3 — the round trip (53% → 43%). The 4.Nd5 sortie was fine, but it needed follow-through: 6.a3 keeps the visit profitable with b4 coming, when the a5-bishop starts running out of squares. Retreating gave back both tempi and the initiative with them.

The invisible pawn move. Three moves running — 12.e4, 13.Rd1, 14.Nd5 — the engine's first choice was the same humble d3: blunt the e-file, connect the position, keep the plus. Every candidate you generated was an active piece move; the quiet structural one never made the list. Development-mode tunnel, chapter and verse — and it cost the entire opening advantage, 61% → 44%, with no blunder anywhere in it.

23.Bg4 — one move too late (48% → 15%). After 22...Nh5 hit the lifted rook, 23.Re3 held everything — and after a later ...b6, Bxh5 even picks up the stray knight. Instead 23.Bg4 attacked the queen and assumed she was obligated. 23...Nxg3 answered the bigger question first: rook and bishop for a knight, resign. A five-second scan of his checks and captures finds it. The clock context counts too: 22.Qd4 ate 95 seconds and left you 0:49 for the critical moment.

What went right

The opening idea was sound: by move 12 you stood +1.2 with the better structure.
No pessimism spiral — you kept making normal moves in a level position under clock pressure. The leak was the missing forcing-scan, not the mood.

Profile check

Both master keys in one game: tunnel vision (three invisible d3s, one assumed queen-move) with the pessimism lever staying quiet. Allowed-fork ledger +1, on a capture you invited to a square a knight was already watching.

What to drill next

The humble-move check, once per move: "which pawn move improves my position?" — especially when everything feels developed.
Before any move that attacks something: his checks and captures first. The queen is never obligated.
Rook lifts need a bodyguard: before a Rg3-style lift, name every knight square that hits the destination. From g3 that's h5 and f5 — one of them was two moves away.

t3gantry 1590 vs toan6ca 1735 0-1

Analysis

♛ Deep reviews

The story of the game

The key moments

Profile check

What to drill next

The story of the game

The key moments

What went right

Profile check

What to drill next

Key moments

Questions on this game