How many games to reach 1000 Elo? For me, 232, and a pinned pawn
How many games to reach 1000 Elo? For me it took 232 rapid games on Chess.com. My first game was on January 10, 2024 at 631; after game #232, on February 28, 2024, my rating read 1001 for the first time. That's seven weeks, 100 wins, 125 losses and 7 draws. The game that pushed me over was a 22-move win with Black against MenkdeTenk (968), and it turned on one pinned pawn.
This page is the story of that game and of the numbers around it. Your number will be different; this is just mine, from my own game history, with every evaluation from Stockfish.
The 232 games in numbers
The honest answer to "how many games" is that most of them were spent standing still. I crossed 900 at game #32. From game #31 to game #214 my rating never went below 846 or above 944. Then, in 18 games, it went from 903 to 1001.
| Stretch | Games | Rating |
|---|---|---|
| Week 1 | #1–#30 | 631 → 875 |
| Weeks 2–7 | #31–#214 | between 846 and 944 the whole time |
| Week 8, up to Wednesday | #215–#232 | 903 → 1001 |
| Total | 232 | 100 wins, 125 losses, 7 draws |
So the jump from the 600s into the high 800s took one week, and the last hundred points took about six weeks of going sideways and then a few good days. The full plateau is in the weekly diaries between those two links, if you want to see what it looked like game by game.
The four games before
Game #232 closed a four-game winning streak, all on February 28. In the morning I beat IgorM1964 (1113) with my best accuracy of the week, 96.7. In the afternoon I played RikardQV (978, then 970) twice, once with each colour, and won both. My rating after game #231 was 993, the highest it had ever been. Then came this game, which ended 18 minutes after the previous one.
| Game | #232, February 28, 2024 |
| Opponent | MenkdeTenk (968) |
| My colour, rating after | Black, 1001 (from 993) |
| Time control | 20 minutes, no increment |
| Opening | King's Pawn Opening, King's Knight Variation (1.e4 e5 2.Nf3 Nc6 3.d3) |
| Result | Win, checkmate on move 22 |
| Accuracy (Stockfish) | Me 89.3, White 78.6 |
| My blunders / mistakes / inaccuracies | 1 / 0 / 3 |
A quiet opening and an open f-file
The opening was calm. White played d3 and a3, I put my bishop on c5 and castled, and on move 7 I pushed my f-pawn two squares. A knight trade on f3 followed, my f-pawn was traded for White's e-pawn, and by move 13 my queen and my rook were both on the f-file, pointing at White's bishop on f3. That bishop was defended by White's queen and by the pawn on g2. Keep that pawn in mind; the game is about it.
Moment 1: a free pawn I didn't take
On move 14 White played the pawn to b3, which the engine called a mistake. White's knight was on d5, and I pushed my c-pawn to chase it away.
Stockfish preferred taking the pawn on h3 with my bishop, the pawn in front of White's king. The tool's description: the bishop can be captured, but the engine's line shows why it still works, and in that line White doesn't take it at all. My pawn move kept the position level (-0.02); the bishop capture would have left me slightly better. A small miss, but the same pawn on h3 will come back later.
Moment 2: the "blunder" that won the game
White's knight went back to c3, and the engine flagged that as a blunder: suddenly Black was winning (-3.83). It wanted my dark-squared bishop to d4, a double attack (one move that threatens two things) on the knight on c3 and the rook on a1 behind it. I played my queen to g6 instead.
Stockfish labels my queen move a blunder because it dropped me from winning to "only" clearly better (-1.69). That's what the label means: the move threw away a lot compared with the best one, not that it lost. And look at what it did on the board. On g6 my queen faces White's king on g1 along the g-file, with the g2 pawn in between. That pawn is now pinned: it can't move without exposing its own king. A pinned pawn can't recapture, so it no longer really defends the bishop on f3. My rook on f8 was now attacking a bishop that had only one real defender left: White's queen on d1.
Moment 3: White's queen steps away
White moved the queen to d2. The tool's description of that move is the whole story: the queen stops defending the bishop on f3, which can now be captured. The engine's choice for White had been king to h2. My rook took the bishop, and the evaluation went to -8.80. On the next moves my rook took the pawn on h3 (there it is again) and then the knight on c3. I was a bishop and a knight up with White's king exposed.
Moment 4: the finish
On move 20 White pushed the pawn to g3. That pawn had a defender on f2, but my bishop on c5 was looking down the diagonal through f2 to White's king on g1, so the f2 pawn was pinned and couldn't take back.
The engine's top move was the rook capture, but it rates both as forced checkmate, so this one doesn't matter. My queen took on g3 with check, White's king went to h1, my queen checked again from h3, the king went back to g1, and my rook came to g3: checkmate on move 22. I still had 14 and a half of my 20 minutes. You can replay the game on Chess.com.
What reaching 1000 actually meant
It would be nice to say I reached 1000 because I had become a better player. The data is less flattering. In this game my opponent made the decisive mistake: the queen move that left the bishop with no defender. I also didn't play the engine's best move at either of the two moments where it mattered most. And across that week I was still making almost two blunders a game.
What I can say is that the game had one idea I'm happy with: I made a threat that wasn't obvious. A bishop that looks defended by a pawn isn't defended if that pawn is pinned. That's the lesson from game #232, and it works in both directions: before I count a defender, I check whether it's free to move.
What changed afterwards
Not much, at first. I lost the next two games and was back at 985 the following morning. The next time my rating read 1000 or more was game #263, a draw on March 5, 31 games later. Week 9 was my first week finishing above 1000, at 1023. So reaching 1000 was less a door I walked through and more a line I crossed, fell back from and crossed again.
If you're counting your own games to 1000, my one piece of advice from this is to count something else too: the moments where your opponent left a piece with no real defender, and whether you saw them. That's the kind of moment I built Chess Coach to find and explain; see whether Claude can analyze your chess games, or follow the whole climb from the journey page.
Frequently asked questions
How many games did it take to reach 1000 Elo?
In my case, 232 rapid games on Chess.com, played between January 10 and February 28, 2024. I started at 631 and my rating read 1001 after game #232.
Did you stay above 1000 after reaching it?
No. I lost the next two games and dropped to 985. The next time my rating read 1000 or more was game #263, a draw on March 5, 2024, 31 games later.
Was the game that took you to 1000 a good game?
Mostly. Stockfish gave me 89.3% accuracy, with three inaccuracies and one move labelled a blunder. That 'blunder' still left me clearly better, and it set up the threat my opponent missed.
Where did most of those 232 games go?
Into a plateau. I passed 900 at game #32, and from game #31 to game #214 my rating stayed between 846 and 944. The last 100 points came in 18 games.