# Game #627: the king and pawn vs king stalemate that saved half a point

> Game #627: a Giuoco Piano where I missed a free rook, gave my bishop for a pawn, and was saved by a king and pawn vs king stalemate on move 73.

- URL: https://thechesscoach.netlify.app/journey/opening-lesson-game-627/
- Updated: 2026-10-28
- Author: Ignacio Soler

On June 30, 2024, in game #627 of my rapid journey, I had Black in a Giuoco Piano against **ECaris (1046)**, and the game ended in a king and pawn vs king stalemate: White's pawn was on b7, White's king on b6, my king on b8, and I had no legal move. That stalemate saved me half a point, because two moves before it I had been lost. And I was only lost because, much earlier, I'd been completely winning and let it go.

The series picked this as a Giuoco Piano lesson. The opening was fine. The real lessons came in the last forty moves.

## Where I was

Game #627 sits in [week 25 of the journey](/journey/week-2024-w26/), an 11–9–2 week that took me from 1020 to 1038. That Sunday evening had started well: I'd beaten Lyan001 (1028) and cyberpunk15 (1021) and was at **1046**. ECaris was rated exactly the same.

The time control was again 10 minutes with no increment, which matters for the long ending that follows.

| | |
|---|---|
| Game | #627, June 30, 2024 |
| Opponent | ECaris (1046) |
| My colour, rating after | Black, 1046 (unchanged) |
| Time control | 10 minutes, no increment |
| Opening | Giuoco Piano Game (C50) |
| Result | Draw by stalemate after White's 73rd move |
| Accuracy (Stockfish) | Me 74.4, White 61.4 |
| My blunders / mistakes / inaccuracies | 9 / 0 / 7 |

## The opening: the knight that went to e7

The moves were 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.O-O d6 5.h3. That's the same position I reached in [game #480](/journey/opening-lesson-game-480/), the game where I missed the Bxf2+ trick. There I played the knight to f6. This time I chose 5...Nge7.

> **Game #627, move 5: White's pawn on e4 has no defender**
> Position (FEN): `r1bqk1nr/ppp2ppp/2np4/2b1p3/2B1P3/5N1P/PPPP1PP1/RNBQ1RK1 b kq - 0 5`
> Played: Nge7 — Nge7, the knight develops behind the bishop's diagonal
> Better: Nf6 — Nf6, the knight develops and attacks the pawn on e4

The engine called 5...Nge7 an inaccuracy (+0.06 to +0.98; evaluations are from White's side, so plus means better for White). Its move was the knight to f6, attacking the e4 pawn, which had no defender. White replied with an inaccuracy of its own, 6.d3, and after 6...Na5 7.Bb3 Nxb3 8.axb3 I had traded my knight for White's Italian bishop, the piece that aims at f7. My opening accuracy was **92.0**. No harm done, but the engine's knight move was f6.

## Moment 1: the pawn push that dropped a piece, and the reply that dropped it back

The middlegame was quiet until move 23. White's bishop sat on e5, my queen on a5, and I pushed my b-pawn.

> **Game #627, move 23: White's bishop on e5 has no defender**
> Position (FEN): `3r2k1/1p3ppp/4b3/qp2B3/8/2P2Q1P/1P3PP1/5RK1 b - - 2 23`
> Played: b4 — b4, the pawn opens the line from my queen to the bishop on e5
> Better: f6 — f6, the pawn attacks the bishop on e5

My move did attack the bishop: the pawn left b5, and my queen on a5 now saw e5. But the engine called it a blunder (+0.07 to +5.87). White's best answer was the queen to g3, which attacked my g7 pawn with the king behind it and threatened mate on the next move. Its move for me was f6, attacking the bishop with a pawn.

White didn't find it. 24.Qxb7 grabbed a pawn and left the bishop on e5 hanging, and I took it: 24...Qxe5, the move the analysis marks as my good moment. The evaluation went to **-5.83**. A few moves later the queens came off, and I had a bishop against two pawns, with a rook each.

## Moment 2: the free rook on g4

From there, the ending was a long chain of chances for me and chances back for White. Between moves 31 and 44 I made six blunders and White made five, plus a mistake; the evaluation swung between clearly winning for me and equal again and again. One square kept coming up: g4. White's rook went there on move 41, came back on move 43, and returned with check on move 46.

> **Game #627, move 46: White's rook gives check from g4**
> Position (FEN): `2b5/8/8/3r2k1/1P4R1/6K1/1P6/8 b - - 2 46`
> Played: Kf5 — Kf5, the king steps out of check and attacks the rook
> Better: Bxg4 — Bxg4, the bishop simply takes the rook

The rook on g4 was checking my king on g5, and my bishop on c8 could take it. The diagonal from c8 to g4 was completely open: d7, e6 and f5 were all empty. Stockfish's verdict on bishop takes g4 was "black can force checkmate". I moved my king to f5 instead, attacking the rook. Look at where that put the king: on f5, the one square of that diagonal that blocks my own bishop. The evaluation went from **-8.69 to -0.13**. I had 73 seconds on my clock.

Two moves earlier I'd made a similar miss: on move 44, with my king on f5 attacking the same rook, the engine wanted a rook check first and then king takes g4. I took a pawn with check instead, and the evaluation went from -6.76 to zero. After 46...Kf5 we traded rooks on move 49, and I was left with my bishop against White's two b-pawns. The engine called that about equal.

## Moment 3: giving the bishop for a pawn

For about fourteen moves the kings and my bishop manoeuvred on the queenside. Then, on move 64, White pushed a pawn to b5, right next to its king on c5, and I took it.

> **Game #627, move 64: White's pawn on b5 is protected by the king on c5**
> Position (FEN): `8/8/b3k3/1PK5/8/1P6/8/8 b - - 0 64`
> Played: Bxb5 — Bxb5, the bishop takes the pawn but can be taken by the king
> Better: Bc8 — Bc8, the bishop steps back and keeps watching the pawns

The pawn was protected. 64...Bxb5 65.Kxb5 left me with a bare king against king and pawn, and the engine's verdict jumped from +0.27 to **+6.32**, "white is completely winning". Bishop to c8 kept the draw. I had 37 seconds; White had 57.

## Moment 4: the pawn push that gave it back

White's king and pawn now needed to win against my lone king. On move 68 the engine's move was the king to b5, keeping the king in front of its pawn.

> **Game #627, move 68: White to move, with king and pawn against my king**
> Position (FEN): `8/1k6/8/2K5/1P6/8/8/8 w - - 3 68`
> Played: b5 — b5, White pushes the pawn
> Better: Kb5 — Kb5, White's king steps in front of its pawn

White pushed the pawn instead, and the evaluation fell from +6.30 to +0.16. From then on my king stayed in front of or next to the pawn, on c7, b7, b8 or c8, and the engine's numbers stayed at zero. On move 72 the pawn reached b7 with check, and after 72...Kb8 White played 73.Kb6.

> **Game #627, final position: Black to move, and Black has no legal move**
> Position (FEN): `1k6/1P6/1K6/8/8/8/8/8 b - - 2 73`

That's stalemate. My king on b8 wasn't in check, but every square around it was covered: a8 and c8 by the pawn on b7, a7 and c7 by White's king, and the pawn itself was protected by the king, so I couldn't take it. A player with no legal move who isn't in check doesn't lose: the game is a draw. The engine's choice on move 73 was the king to d6, which also scored 0.00, because then my king could take the pawn. By that point White had no win left, so the stalemate only ended a drawn position sooner.

## What the engine says about game #627

My accuracy was **92.0 in the opening, 78.7 in the middlegame and 70.0 in the ending**. Nine blunders, seven inaccuracies, no mistakes. White made eight blunders, one mistake and three inaccuracies. My best winning chance during the game, by the engine's estimate, was **96.1%**. I finished with 26.6 seconds on my clock; White had 46.

## The lesson I took from game #627

The king and pawn vs king stalemate is the defender's best friend, and this time I was the defender. The final position is worth memorising: when the pawn reaches the seventh rank with check and the defending king is in the corner in front of it, the wrong king move by the stronger side leaves no legal moves at all. In this game it didn't change the result, because the position was already drawn, but it's the pattern that saves you when it isn't.

The bigger lesson is about captures. On move 46 a whole rook was free, and I moved my king onto the one square that blocked my own bishop. When a piece of mine gives check or is checked, I now look first at what can capture the checking piece, before I think about king moves.

And the last one: in a bishop against pawns ending, the bishop is the reason it's a draw. Giving it for one protected pawn turned a level position into a lost one in one move.

## What came next

My next game was a loss to brodieruss (1062) in an Englund Gambit, and I finished [week 25](/journey/week-2024-w26/) at 1038. The week before, [game #588](/journey/opening-lesson-game-588/) had ended in a draw too, with my king and rook unable to mate a lone king in time. The following week, [week 26](/journey/week-2024-w27/), took me to a new high. The whole series is on [the journey page](/journey/), and you can [replay game #627 on Chess.com](https://www.chess.com/game/live/108015791604).

Chess Coach finds moments like 46...Kf5 in your own finished games, draws the better move on a board, and explains in plain words why it was better, inside Claude or ChatGPT: here's [how the analysis works](/features/ai-chess-analysis/).

## FAQ

### What is stalemate in a king and pawn vs king ending?

Stalemate is when the side to move is not in check but has no legal move, and the game is a draw. In game #627 my king was on b8, White's pawn on b7 and White's king on b6. The pawn covered a8 and c8, the king covered a7 and c7 and protected the pawn, so my king had no move at all.

### Was the king and pawn ending in game #627 winning for White?

For a few moves, yes. After I gave my bishop for a pawn on move 64, Stockfish rated the position about +6.3 for White. On move 68 White pushed the pawn to b5 instead of bringing the king to b5, and the evaluation fell to about equal. The stalemate on move 73 came after that.

### What was the biggest mistake in game #627?

On move 46 White's rook gave check from g4, and my bishop on c8 could simply take it. Stockfish rated that capture as a forced mate for me. I moved my king to f5 instead, which blocked my own bishop's diagonal, and the evaluation went from -8.69 to -0.13.

## Related

- [From 631 to 1385: my chess journey in 2,117 games](https://thechesscoach.netlify.app/journey.md)
- [Week 25 (Jun 2024): I missed a mate in 3 and it cost the game](https://thechesscoach.netlify.app/journey/week-2024-w26.md)
- [The Bxf2+ trick in the Italian Game that I didn't play](https://thechesscoach.netlify.app/journey/opening-lesson-game-480.md)
- [How to checkmate with king and rook: my 28-second draw](https://thechesscoach.netlify.app/journey/opening-lesson-game-588.md)