Reaching 900: how long it took me to get to a 900 chess rating
How long does it take to get to a 900 chess rating? For me it took 32 rapid games over six days, from my first rated game on January 10, 2024 to a Monday morning, January 15. The rating Chess.com recorded with my games read 892 at game #31 and 908 at game #32, a win with Black against debabrata882, rated 1041. The first 800 had come at game #3, so the second hundred points took about ten times as many games. And as I found out within the week, reaching 900 and staying there are not the same thing.
How long it took me to get to 900, step by step
| Milestone | Game | Date (2024) | My rating in the game record |
|---|---|---|---|
| Start | #1 | Jan 10 | 631 |
| First time at 800+ | #3 | Jan 10 | 868 |
| End of week 1 | #30 | Jan 14 | 875 |
| First time at 900+ | #32 | Jan 15 | 908 |
| Week high | #34 and #36 | Jan 15 | 931 |
| After nine losses in a row | #45 | Jan 17 | 871 |
| First time at 1000+ | #232 | Feb 28 | 1001 |
Between 800 and 900 the number went up and down a lot. In my first week I touched 897 at game #9 and 896 at game #11 and then fell back to 804 by game #21. The typical move from one game to the next in those games was 22 points (the median). So 900 wasn't a straight climb: I had come within a few points of it twice before, and fallen back both times.
The game: debabrata882 (1041) vs me
Game #32 was the second of three 15|10 games I played that Monday morning (fifteen minutes each, plus ten seconds a move). It started at 09:27 UTC and finished at 09:47. My opponent opened with the king's fianchetto, pushing the g-pawn one square and putting the bishop on g2; Chess.com files it as "King's Fianchetto Opening". I answered with pawns to e5 and d5 and quick development.
Stockfish rates it as one of my cleaner games from that period, though not a clean one:
| debabrata882 (White) | Me (Black) | |
|---|---|---|
| Rating | 1041 | 908 |
| Accuracy | 82.3% | 89.4% |
| Inaccuracies | 4 | 2 |
| Mistakes | 0 | 1 |
| Blunders | 2 | 1 |
Evaluations below are from White's side, so here a minus sign is good for me: -2.00 means Black is about two pawns better.
Move 9: White's pawn push, and a pawn I could take in two ways
On move 8 I pushed my d-pawn to d4 (the engine calls that an inaccuracy). White answered 9.e4, a blunder according to Stockfish, which wanted the pawn on d4 taken instead. I took on c3, which was the engine's move too, and after 10.Nb3, attacking my bishop on c5, I was at -3.98. Then came my own mistake of the game.
I thought for 63 seconds and took on b2. That pawn attacked the rook on a1 and the bishop on c1 at once, which looks great, but White simply took it back with the bishop. The engine's move was to push the pawn to c2 instead, attacking the queen on d1. In its line White's queen steps to d2, my bishop comes to b4 to attack the queen with the king behind it, and White's knight goes to c3. My capture brought the evaluation from -3.98 to -2.06. Still clearly better for me, but I had given back half of what White had handed me.
Move 15: the knight square that let my opponent back in
Five moves later I was at -2.74 and chose to put my knight on b4, after 41 seconds. Stockfish marks it as my only blunder of the game, and looking at it now, the problem is visible in one look.
On b4 my knight had no defender. White's queen could go to a4 and, from that one square, attack both the loose knight and my bishop on a6. That's a double attack, and after Nb4 the engine rated the position -0.10, about equal. It preferred the knight on d4, in the center, where it keeps Black clearly better (when I re-checked the position, the engine also liked bringing my queen to d4).
Move 16: my opponent's answer
debabrata882 didn't play the queen to a4. White played 16.Nd5 instead, a knight jump that attacked my knight on b4 and my other knight on f6.
I took it with the knight from b4, which was the engine's first choice, and after the pawn took back I collected that pawn as well with my other knight. The evaluation went to -5.49 after White's knight jump and never came back above -3.49 for the rest of the game. One move on each side decided it: mine let White back to equal, and White's gave it all back.
The rest: a long squeeze, and one more thing I missed
From there it was a slow game. After a few trades each of us had a queen and a rook, White a knight and me a bishop. I picked up the a2 pawn with my queen on move 27 and the d3 pawn with my bishop on move 32, then traded my bishop for White's knight and pushed my e-pawn down to e3, next to my pawn on d4. On move 36 White put the rook on a2, where my queen on d5 could take it with check. The engine's line for that position is exactly that capture. I played my a-pawn forward instead, in 11 seconds. My opponent didn't make another move: the game record says "nacsol3 won - game abandoned". You can replay it on Chess.com.
One more number from that ending: after move 36 I had 8:01 on my clock and White had 15:14. I'd spent almost twice as much time and still missed a free rook at the end.
The lesson: undefended pieces are where it goes wrong
Both my worst move (15...Nb4) and the move I missed at the end (36...Qxa2+) were about the same thing: a piece with no defender. On move 15 it was my knight, landing on a square where the opponent's queen could attack it together with another loose piece. On move 36 it was White's rook, sitting where my queen could take it with check. In between, White's 16.Nd5 put a second knight in front of my pieces and lost material for it.
The question I took from this game is simple: after my move, which of my pieces has no defender, and can one enemy move hit two of them? On move 15 that question has an answer in one look: the knight on b4 and the bishop on a6 can both be reached by the queen from a4.
What changed after reaching 900
The good run lasted one more game. I won game #33 (920), the last of my three 15|10 games that Monday, and then moved to 10|5. I was at 931 after games #34 and #36, and then lost nine in a row, from game #37 to #45, down to 871. That stretch is in my week 2 diary: those three 15|10 games, this one included, were my only games of the week at that time control, and I won all three. I finished the week at 899, one point under the number this post is about.
In the 199 games between #33 and #231, my rating was below 900 in 117 of them, and the lowest point was 846 at game #172, on February 11. The 1000 mark came at game #232, on February 28. So if you're asking how long it takes to get to 900, my honest answer is: 32 games to touch it, and about 200 more to be done with the 800s. The later climb from 900 is in week 26, the road from 900 to 1200.
What I'd change looking back is how I used games like this one. A win against a 1041 felt like proof, and on the result alone it was. The engine shows two knight moves, mine and White's, deciding the game, and a rook left on the table at the end. If you want to check your own games like this, here is how Claude can analyze chess games with Stockfish behind it. The full story, from 631 to 1385, is on the journey page.
Frequently asked questions
How long does it take to get to a 900 chess rating?
For me it took 32 rapid games on Chess.com, from January 10 to January 15, 2024, starting at 631. Staying there was harder: in the next 199 games my rating was below 900 in 117 of them.
Is 900 a big jump from 800?
In my case the first 800 came at game #3 and the first 900 at game #32, so the second hundred points took about ten times as many games. The first 1000 came at game #232.
Did you win the game that took you to 900 cleanly?
No. I was clearly better, then put a knight on b4 where it could be attacked together with my bishop. My opponent, rated 1041, missed it and played a knight move that lost material instead.