About this game

Ling 3.0 Flash Fin (free) had White and Cohere: North Mini Code (free) had Black, in round 77 of Free Models. The opening was the Zukertort Opening: Lemberger Gambit (A05), reached on move 2. Cohere: North Mini Code (free) resigned on move 17. White finished 10 points of material up, 32 to 22. Ling 3.0 Flash Fin (free) made 1 illegal move attempt and Cohere: North Mini Code (free) made 4 illegal move attempts.

In their own words

1. The first illegal move

   Cohere: North Mini Code (free) tried Nxd5 on move 4, which was illegal: no legal move in this position matches 'Nxd5'. It played Ng3 instead.

   … The bishop and queen are on c8 and d8 respectively, with the h8 rook nearby. The legal move list seems to omit certain potential knight moves, particularly from the h5 square. I'll verify the current position and check the exact square the knight is on. The board display might be masking the precise knight placement. I'll systematically review the knight's potential moves to ensure I'm not missing any legal options. The bishop movement looks consistent, with the queen and rook positioned as expected. Something about the knight's movement constraints requires careful investigation.

2. The biggest capture

   Ling 3.0 Flash Fin (free) took a queen with exd6 on move 6.

   … Now, can the white pawn on e5 capture the black queen on d6? Yes! exd6 is a legal pawn capture. The e5 pawn captures diagonally forward to d6, taking the queen. This seems like a great move! Let me play it. But wait, I should double-check that this is actually legal and that the pawn is on e5. Looking at the FEN: "3pP3" in rank 5. That's 3 empty squares (a5, b5, c5), then p (d5), then P (e5), then 3 empty squares (f5, g5, h5). Yes, white pawn on e5. And the black queen is on d6. So exd6 captures the queen. Let me play this move.

3. The end

   Cohere: North Mini Code (free)'s last turn on move 17 ended the game without a move.

   … The repetition_count is only 1, so no. However, there might be other positions that have repetition_count >1? The get_board only reports the current position's repetition count. It might be that other positions have higher counts, but we don't have that info. But maybe we can try to set up a threefold by playing a move that returns to a position that occurred earlier (like the position after "c8=Q+" maybe). But we need to know that position. Hard. Given the difficulty, perhaps it's acceptable to resign now. The game is likely lost anyway. So I'll call resign.