♞ ChessBay
♞David ♝James

t3gantry 1643 vs alihaider007 1575 0-1

2026-07-20 English Opening: Agincourt Defense 10+5 rapid lichess ↗
alihaider007 1575 1
Start

Tap any move, use ← → keys, or drag pieces to try your own ideas.

t3gantry played white 2 blunders 1 inaccuracy turned on +3.1 → -3.1

Analysis

Nine pawn moves, one won game, two autopilot moments

The story of the game

James, you met the crawl exactly right. Black's first nine moves were all pawn moves — ...e6, ...c6, ...g6, ...a6, ...h6, ...b5, ...b4, ...a5, ...d5 — and your answer was the classical refutation: the full centre with 4.d4 and 5.e4, pieces out, 12.O-O. The holes the pawns left behind became material on schedule: 15.Nb6 jumped into the queenside, 16.Bf4 ran the rook out of squares, and 16...Rxb6 17.cxb6 banked the exchange. 19.Bxg6+ added a pawn, and by move 25 the engine had you at 91%. What lost the game wasn't the opening or the middlegame — it was two moves played on autopilot in a "finished" position: the automatic recapture 28.Rxc1, and the queen trade 32.Qe3 offered while your queen was the only defender of the f7-bishop. After 32...Qxe3 33.Rxe3 Nxf7 the two knights plus the queenside pawn mass beat the rook.

The key moments

  • 15.Nb6! — the crawl's bill comes due. The knight that got pushed to the rim on move 8 forks into the squares ...b4/...a5 abandoned. 15...Rb8 16.Bf4 hit the rook again, and 16...Rxb6 17.cxb6 won the exchange. Model punishment: not a refutation hunt, just development pointing at holes.
  • 25.Nxd7 (91% → 80%). The engine wanted Qg6, threatening to collect the h6-bishop: after 25...Nfe4 26.Qxh6 White is +5.7 and Black's king has no shelter. 25.Nxd7 swapped your best attacker — the e5-knight — for the passive bishop. Still winning, but the bind loosened. Keep the attackers; trade the bystanders.
  • 28.Rxc1?? (81% → 54%). The bishop on c1 wasn't going anywhere. Bxe6! first — grabbing a pawn, hitting d5, eyeing the d7-knight — keeps every threat alive at +3.9, and c1 falls next move anyway. The instant recapture let 28...Qxd4 centralize with tempo, and suddenly the queen-and-knights swarm was real.
  • 32.Qe3?? (69% → 26%). The same idea one tempo out of order. Bxe6! saves the attacked bishop with profit, and then the queen check on e3 trades into a winning ending. Played immediately, the trade left f7 undefended: 32...Qxe3 33.Rxe3 Nxf7 and the piece was simply gone.

Phase notes

  • Opening: keep doing exactly this. Centre, development, castle — you were strategically winning by move 12 without a single flagged move.
  • Middlegame: the conversion habit to build is treating recaptures and trades as real decisions. Both blunders were "obvious" moves; neither survived a five-second captures-checks-threats scan.
  • Endgame: by the time the R-vs-2N ending arrived it was lost — the knights had targets everywhere. Worth replaying move 36 onward once to feel how fast connected passers plus knights outrun a rook.

Profile check

Three profile items fired at once: the captures-checks-threats scan collapsing exactly when the position feels decided; the relocated-pieces alarm, extended to trades — Qf4 defended f7, Qe3 didn't; and "simplify when ahead," which needs its missing clause: after checking what the trade un-defends. New fork-family entry: the zwischenzug-before-recapture check.

What to drill next

  1. Winning-position discipline: in your next three won games, do the full scan on every capture and every offered trade — count it a win only if you convert without a 20-point swing.
  2. Zwischenzug puzzles (in-between moves before recaptures) — 10 a day for a week.
  3. Replay this game from move 27, playing White's side against the engine, converting with Bxe6 ideas.

♛ Deep review

for ♝ James Claude Fable 5 2026-07-20

The story of the game

James — I know this opponent's pawn-crawling drove you up the wall, so let's start with the verdict your annoyance was fishing for: yes, it's bad chess. Nine pawn moves in the first nine turns is objectively terrible, and your response was the textbook one — 4.d4, 5.e4, develop, 12.O-O. By move 25 the engine scored you at 91%. You won the argument about the opening. The game was lost somewhere else entirely: in a position you'd already decided was over, with two autopilot moves — 28.Rxc1 and 32.Qe3 — that each failed a scan you normally run. That's the real content of this loss. Annoyance doesn't blunder for you on move 10; it waits until move 28, when the opponent has been "beaten" and your attention has already left the board.

The key moments

15.Nb6! — remember how this win was built. No refutation hunt, no lunge. You developed, castled, took the centre, and the knight he chased to the rim on move 8 landed in the exact hole ...b4/...a5 created. 15...Rb8 16.Bf4 ran the rook down, 16...Rxb6 17.cxb6 won the exchange, 19.Bxg6+ took a pawn with check. This is the whole anti-crawl recipe: the punishment is a won middlegame, delivered by boring moves.

25.Nxd7 — don't cash the attackers (91% → 80%). Qg6 instead threatens the h6-bishop, and after 25...Nfe4 26.Qxh6 it's +5.7 with his king naked. Your e5-knight and queen were the attack; the d7-bishop was a spectator. Trade bystanders, keep attackers.

28.Rxc1?? — the recapture that wasn't forced (81% → 54%). Ask what the bishop on c1 was going to do — nothing. Bxe6! first: a pawn, a hit on d5, the d7-knight suddenly loose, and c1 still falls next move (+3.9). After the instant 28.Rxc1, black played 28...Qxd4 — centralizing with tempo — and his queen-plus-knights machine was back in business. You have the Qa4+ zwischenzug note in your English pocket already; this is the same idea from the other side. Recaptures are moves. They get the same scan as everything else.

32.Qe3?? — the queen was the glue (69% → 26%). The most instructive moment of the pair, because the idea was right and the order lost. Your f4-queen was doing two jobs: eyeing e5, and defending f7 down the file. Bxe6! first — the attacked bishop escapes with a pawn and tempo — and then Qe3+ next move trades queens into a winning ending. Same trade, one tempo later, totally different game. Played immediately: 32...Qxe3 33.Rxe3 Nxf7, piece gone, and the 2N + queenside pawns ending is lost for the rook. Before offering any trade: what is this piece currently defending?

Phase notes

  • Opening: flawless by engine count — zero flags before move 25. Bottle this: vs system crawls you now have proof the boring recipe wins.
  • Middlegame: the leak is specifically conversion moves — recaptures, trades, "cleanup." Those got a different, lazier mental process than your attacking moves. They need the same one.
  • Endgame: after 33...Nxf7 the R-vs-2N ending was gone, but note how you kept creating problems until move 42 — that fighting habit is worth keeping even when it doesn't save the game.

Profile check

This game is your Mode-Locked Tunnel Vision master key wearing a new costume: the scan died the moment the game felt won. It also writes a new clause into "simplify when ahead" — simplify after the trade-safety check. And the relocated-pieces alarm now covers trades: re-read what a piece defends before you offer to remove it from the board.

What to drill next

  1. The winner's scan: next three winning positions you reach, run captures-checks-threats on every single move until resignation — especially recaptures. The scan you run at move 15 has to survive to move 35.
  2. Zwischenzug set: a week of in-between-move puzzles; you're missing them in one specific spot — before recaptures.
  3. Replay from move 27 against the engine as White. Find Bxe6 both times. Feel the difference between "the game is over" and "the game is over after I play accurately."

Key moments

Questions on this game

Nothing asked yet.