Game #480: the Italian Game Bxf2+ trick I missed against a 1218
On May 28, 2024, in game #480 of my rapid journey, I had Black in a Giuoco Piano against admcastr (1218), and on move 7 White's rook move to e1 left the Bxf2+ trick on the board: bishop takes the f2 pawn with check, then takes the rook. Stockfish scored it as completely winning. I moved my knight back instead. Later I got a winning position again, gave it back, and resigned on move 26 in a position the engine calls only slightly better for White.
My rating dropped to 999. It was the first time I'd been under 1000 since game #424.
The series picked this one as a Giuoco Piano lesson, and it's the right pick: the opening gave me a free rook, and I didn't take it.
Where I was
Game #480 sits in week 21 of the journey, the week I went 5–14 and slid from 1026 to 975. That Tuesday was rough. Before this game I'd lost to Ya7yoot (1128) and waljara116 (996), beaten khaledkhaled67 (993), and lost the rematch with Black. I came into game #480 at 1003, against someone 215 points higher.
The Giuoco Piano with Black was a problem for me at the time. Counting this game, my record as Black in games Chess.com labels simply "Giuoco Piano Game" was 3 wins, 11 losses and 1 draw.
| Game | #480, May 28, 2024 |
| Opponent | admcastr (1218) |
| My colour, rating after | Black, 999 (from 1003) |
| Time control | 15 minutes + 10 seconds per move |
| Opening | Giuoco Piano Game (C50) |
| Result | Loss, I resigned after White's 26th move |
| Accuracy (Stockfish) | Me 64.5, White 67.8 |
| My blunders / mistakes / inaccuracies | 7 / 1 / 2 |
The opening: a free pawn, then a free rook
The moves were 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.O-O d6 5.h3 Nf6 6.c3. White's c-pawn push was an inaccuracy by the engine's count: its choice was the rook to e1, and after c3 nothing guarded the pawn on e4. I took it, 6...Nxe4, and that was the engine's move too.
Then White played 7.Re1, attacking my knight on e4 along the file. The engine called it a blunder: -0.48 to -6.99 (evaluations are from White's side, so minus means better for me).
Here's the trick. With the rook gone from f1, the pawn on f2 was defended only by the king. Bishop takes f2 was check, and it also hit the rook on e1: a double attack (one move that threatens two things at once). When I checked the position again for this post, Stockfish's line was 7...Bxf2+ 8.Kh2 Bg3+ 9.Kg1 Bxe1 10.Qxe1, at -7.35, "black is completely winning". Its second choice was knight takes f2, attacking White's queen, also completely winning.
I saw that my knight was attacked and moved it. The evaluation went to +0.53. Of all the moves in this game, this is the one I'd most like back, because the threat against me and the win for me were on the same file. The rook attacking my knight was the same rook the bishop could take.
Moment 1: the pawn push I didn't make
White kept giving chances. 8.Qb3 was a blunder (+0.53 to -2.22), I castled, and 9.Ng5 was another (-2.68 to -5.56). White's queen on b3 and knight on g5 were both aiming at f7, next to my king.
I defended with the queen on e7. The engine's move was d5: a pawn push that blocks the bishop's diagonal to f7 and attacks the bishop at the same time. My move kept me better (-2.47) but gave back half of the edge. Looking at it now, I defended against the threat when I could have attacked the piece making it.
Moment 2: the bishop I didn't take
By move 13 my knight had reached a5 and White's bishop was still on c4. White's queen came to a4.
Knight takes the bishop on c4 was the engine's move, at -4.58. I developed my bishop to d7 instead, attacking the queen, and the engine logged it as a blunder: -4.55 to -0.36. White returned the favour right away with 14.Bb5 (-5.42), I answered 14...c6, which was the engine's move, and the bishop went back to c4. Then I blundered again with 15...d5 and recaptured on f6 with the queen instead of the g-pawn on move 16. White gave it all back once more with 17.Bxb6. That stretch was a tennis match.
Moment 3: the discovered attack
By move 21 I was clearly winning again, at -6.23. White's queen was on a3, on the same file as my rook on a8, with only my knight on a5 between them.
Knight to c4 was a discovered attack: the knight moves out of the way, and the piece behind it (my rook on a8) suddenly attacks something (the queen on a3). The knight on c4 also attacked the queen itself, the rook on e3 and the knight on d2. The engine's line was 22.Nxc4 Rxa3, winning the queen for a rook and knight, still "completely winning". I played the king into the corner, and the evaluation dropped to -1.91.
Moment 4: the pawn push and the resignation
I pushed the h-pawn to attack the rook on g3. White's queen on b4 took it with check, and the evaluation went from -2.62 to +0.74. I resigned right there.
When I put the final position into the engine for this post, it scored it +0.89, "slightly better for white". Material was level: queen, two rooks, a bishop, a knight and seven pawns each. I had 13 minutes on my clock, plus 10 seconds a move. There's no engine number for why I resigned. The record just shows a player who'd been winning three times, had just lost a pawn with check, and stopped. I wrote a whole post about resigning too early from a different game; this one belongs in the same folder.
What the engine says about game #480
My accuracy was 68.5 in the opening and 60.8 in the middlegame. Seven blunders, one mistake, two inaccuracies. White made five blunders and six inaccuracies. My best winning chance during the game, by the engine's estimate, was 92.9%. The game lasted about twelve minutes.
The lesson I took from game #480
The Bxf2+ trick in the Italian Game comes down to one check: when White's rook leaves f1, look at f2. In this game my bishop had been on c5 since move 3, aiming at that square, and the rook move that attacked my knight was the same move that took f2's second defender away.
More generally, three of my big misses had the same shape. On moves 7, 13 and 21, I had a capture or a knight jump that attacked several things at once, and each time I played something slower. Before I defend now, I try to ask whether I have a capture or a check that's bigger than the threat.
And the last lesson isn't about chess moves: equal positions with 13 minutes left aren't resignations.
What came next
That evening I lost to VanExel (1045) and dropped to 992. I finished the week at 975, the story told in week 21. The next week, week 22, went 11–17–1. The whole series is on the journey page, and you can replay game #480 on Chess.com.
If you've ever looked at a review that marks a move like 7...Nf6 as a blunder without saying why, I wrote about Chess.com's game review not explaining moves. Chess Coach shows the move, the better one and the reason in plain words, inside Claude or ChatGPT: here's how the analysis works.
Frequently asked questions
What is the Bxf2+ trick in the Italian Game?
In my game, after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 my bishop sat on c5 aiming at f2. When White moved the rook from f1 to e1, the pawn on f2 lost a defender, and bishop takes f2 with check also attacked the rook on e1. After the king moves, the bishop takes the rook. Stockfish scored it -7.35, completely winning for Black.
Why didn't you play Bxf2+?
White's rook on e1 was attacking my knight on e4, and I moved the knight back to f6. That threw the advantage away: the evaluation went from -6.99 to +0.53. The engine also liked knight takes f2, so I had two winning captures on f2 and chose neither.
Did you resign a lost position in game #480?
No. I resigned after White's 26th move, with 13 minutes on my clock, in a position Stockfish rates +0.89, only slightly better for White. Material was level.