Reaching 800: how long it took me to get to an 800 chess rating
How long does it take to get to an 800 chess rating? For me it took three rapid games, all on the evening of January 10, 2024. The rating Chess.com recorded with my games read 631 at game #1, 781 at game #2 and 868 at game #3. Game #1 ended at 16:49 UTC and game #3 at 19:33 UTC. This page is about that third game against GoodForkingLord (750): the game that put me over 800, and that, looked at with Stockfish today, I didn't really deserve to win.
How long it took me to get to 800, in numbers
| Game | Date (2024) | Opponent (rating) | Result | My rating in the game record |
|---|---|---|---|---|
| #1 | Jan 10 | mohammad1347reza (488) | Win | 631 |
| #2 | Jan 10 | jon_maik (697) | Win | 781 |
| #3 | Jan 10 | GoodForkingLord (750) | Win | 868 |
| #5 | Jan 11 | Janner_Sasuke (945) | Loss | 771 (the only time below 800 again) |
| #32 | Jan 15 | debabrata882 (1041) | Win | 908 (first time at 900) |
| #232 | Feb 28 | MenkdeTenk (968) | Win | 1001 (first time at 1000) |
So the honest answer has two halves. Getting to 800 was fast: two steps, +150 and then +87. Staying around 800 and climbing beyond it was slow. In games #4 to #30 the typical change from one game to the next was 22 points (the median), and from game #31 to #66 it was 8. The first few games moved my number in much bigger steps than anything that came later, so I don't read "three games" as proof that I already played like an 800. The rest of this post is the evidence.
The game that took me past 800
Game #3 was a 15|10 rapid game (fifteen minutes each, plus ten seconds per move). I had White against GoodForkingLord, rated 750, and the opening was a Queen's Gambit Accepted, the line Chess.com labels "Central Variation, Alekhine System". It lasted 26 moves and ended when my opponent resigned.
Stockfish's summary of the game:
| Me (White) | GoodForkingLord (Black) | |
|---|---|---|
| Accuracy | 67.4% | 60.9% |
| Inaccuracies | 2 | 3 |
| Mistakes | 1 | 2 |
| Blunders | 3 | 4 |
In the opening phase my accuracy was 59.9%, with three errors in twelve moves. That is not the profile of someone who has "solved" 800. The engine's evaluation curve swings from one side to the other several times before move 22. Evaluations below are from White's side, my side here: +1.00 means I'm about a pawn better, a minus sign means Black is better.
The whole game turned around one piece: Black's queen. Three times it was the target, and three times I handled it badly. The fourth time a piece was hanging, I finally took it.
Move 10: I attacked the queen and forgot about my bishop
At +0.96 I jumped my knight from c3 to b5 to hit the black queen on d6. I spent 32 seconds on it. The engine's choice was simply to castle. My knight move allowed the queen to go to b4 with check, and from b4 it would also attack my bishop on c4, which nothing defended. Stockfish's line for Black is the check, then the queen takes that bishop. The evaluation after my move: -6.45.
I wrote about this moment in my week 1 diary. What I didn't say there is what happened next: GoodForkingLord didn't find the check. The queen went to c5 instead, and the evaluation came back to +1.77 for me. The game didn't punish my mistake, and that's exactly why a win like this can teach you the wrong thing.
Move 12: the bishop move that attacked two things
Two moves later Black castled, and the engine rated the position +2.71 for me. I developed my bishop to e3, attacking the queen on c5. That was a natural move, played in 8 seconds. It dropped the evaluation to +0.35. The engine's move was the same bishop going to a3 instead.
From a3 the bishop attacks the queen on c5, and behind the queen, on the same diagonal, the black bishop on e7. That's a double attack: one move threatening two things at once. The engine's line only goes as far as the queen stepping away to b6, and keeps me clearly better. From e3 my bishop only attacked the queen, and the queen answered with a check on b4. Both moves "attack the queen". Only one of them asks what is standing behind it.
Move 16: the free queen I didn't take
This is the moment I find hardest to look at. After 15...Qd6, Black's queen stood on d6, right where my knight on b5 could take it. Stockfish rated the position +7.92: completely winning for me. The capture was there. Black would get my knight back by taking on f3 and then on d6, and I'd still be a queen for a knight up.
I took 46 seconds and played my knight back to d4. The evaluation fell to -0.24, about equal. The engine has no special word for it: it marks it as a blunder, the same label as hanging a piece. I had spent the first fifteen moves chasing that queen, and when it finally stood on a square my knight could reach, I moved the knight away from it.
Move 22: the one I did see
The game went on. My knight went to f5 and traded itself for Black's dark-squared bishop, I pushed my d-pawn, and after a few trades my queen landed on d5. From there it was already looking along the row at Black's bishop on h5. Black answered with 21...Nd7, developing a knight but not dealing with the bishop. The engine had preferred trading that bishop on f3 first.
This time I took it in 6 seconds. The evaluation went to +6.47, and from there it stayed above +5.6. Four moves later Black's knight went to d2, my bishop took it, and GoodForkingLord resigned. You can replay the full game on Chess.com.
The lesson: the queen was never the point
Put the four moments side by side and a pattern shows up. When I attacked the queen (moves 10 and 12), I only looked at the queen. When the queen was actually hanging (move 16), I didn't see it. When a bishop was hanging (move 22), I saw it immediately. So it wasn't that I couldn't see free pieces. I was looking at the board through my plan ("chase the queen") instead of looking at what was actually there each move.
The habit I took from this game is short enough to use in a 15|10 game. Before every move:
- What did my opponent's last move leave undefended? (Move 16: their queen.)
- If my move attacks something, what else is on the same line? (Move 12: the bishop behind the queen.)
- What are their checks after my move? (Move 10: the queen check on b4.)
None of these needs calculation beyond one move. All three would have changed this game.
What changed after reaching 800
The next morning I lost two games in a row and dropped to 771 at game #5. It's the only time in the whole record that I went below 800 again. The rest of the first week bounced between 804 and 897, and I finished it at 875 with 15 wins, 14 losses and a draw, as I describe in week 1. The first 900 came at game #32, five days after this game, and 1000 took until game #232 at the end of February. The slower climb from 800 is the subject of my week 5 diary on going from 800 to 1200.
What changed in how I review games is the other half. Looking only at the result, game #3 was a clean win that pushed me past 800. With the engine it's a game where I was completely winning at move 16, about equal a move later, and won because my opponent left a bishop hanging. If a game review only tells you "blunder" without saying what you missed, it's easy to skip past moments like move 16. That gap is why I wrote about what to do when Chess.com's game review doesn't explain the moves.
The rest of the story, from 631 to my peak of 1385, is on the journey page.
Frequently asked questions
How long does it take to get to an 800 chess rating?
It depends on where you start. I started at 631 on Chess.com rapid, and the rating recorded with my third game, on January 10, 2024, was 868. Getting from there to 900 took until game #32, and to 1000 until game #232.
Did you stay above 800 after reaching it?
Almost. In my next 2,113 recorded games the number dropped below 800 only once: 771 at game #5, the morning after.
Does reaching 800 mean you play well?
Not on its own. In game #3, the one that took me past 800, Stockfish gave me 67.4% accuracy and counted three blunders. I won because my opponent made four.