Week 1: 631 → 875, and how to analyze chess games as a beginner

My first week of rated rapid chess on Chess.com ran from January 10 to 14, 2024: 30 games, 15 wins, 14 losses and 1 draw, and my rating went from 631 to 875 (+244). Most of that jump came straight away (I was at 868 after game #3); after that the number bounced between 771 and 897. Going back through all 30 games with Stockfish is also my honest answer to how to analyze chess games as a beginner: don't replay every move, find the few moments where the evaluation jumps, and ask what you missed right there.

The week in numbers

Every game was 15|10: fifteen minutes each, plus ten seconds per move. As White I opened with the queen's pawn, and ten of my fourteen White games are labelled as a Queen's Gambit line, Accepted or Declined. I won 4 of those and lost 6. As Black, 12 of my 16 games started with 1.e4 e5, and the Italian Game was the one that hurt: 1 win, 4 losses. The engine counted 53 blunders. In 16 of the 25 games where I blundered at least once, the first one came in the opening phase, before I had even finished developing.

The table hides one more thing. In four of my 14 losses (games #7, #10, #21 and #30) the engine had given me at least an 80% winning chance at some point. In game #7 against mabelmoreno44 it was 96.7%, and I still got checkmated.

Three moments that explain the week

The analysis flags the moves with the biggest evaluation swings; these three teach the most. Evaluations are always from White's side: +1.00 means White is about a pawn better, a minus sign means Black is better.

Game #3: a knight attack that walks into a queen check

Against GoodForkingLord (750) I had White in a Queen's Gambit Accepted and the engine had me at +0.96, a small but real edge. I jumped my knight to b5 to attack the black queen on d6. It felt active. What I didn't look at was the answer: the queen can go to b4 with check, and from there it also attacks my bishop on c4 (which had no defender) and the knight on b5. That's a double attack, one move threatening two things at once, and Stockfish rates the position after my knight move at -6.45.

r n . . k . . r
p p p . b p p p
. . . q . n . .
. . . P . . . b
. . B . . . . .
. . N . . N . P
P P . . . P P .
R . B Q K . . R

The engine's move is the boring one: castle. King to safety, rook into play, and the edge stays. My opponent replied ...Qc5 instead of the check, and I won the game, but I didn't earn that one.

Game #4: the "queen trade" that was a free queen

The next morning, with Black against ChocoteBABY (986), I was about a pawn better (-1.17) in an Italian Game. I took the pawn on b3 with my queen. From b3 my queen attacked White's queen on d1 and the bishop on c3, so in my head I was offering a trade on my terms. There was no trade. Nothing protected my queen on b3, so White simply took it, and the evaluation went to +8.87, completely winning for White.

r . . . k . . r
. . p . . . p p
. . n . q p . .
. p . . p . . b
. . . . . . . .
P P B P . N . P
. . . . . P P .
R . . Q R . K .

Castling was the best move again. White took the queen and I resigned on the spot.

Game #5: the pin that wasn't

An hour later, Black again, against Janner_Sasuke (945) in the Italian Two Knights. I played my bishop to g4, attacking the knight on f3 with White's queen behind it. A pin is when a piece can't move without exposing something more valuable behind it, and I thought the knight was stuck. It wasn't: a knight pinned to a queen (not to a king) can still move if it wins something. The engine's answer is knight takes my pawn on e5, and from there it attacks my bishop on g4 and my knight on c6, while White's queen, no longer blocked, also hits the bishop. The evaluation went from +1.00 to +3.41.

r . b q k b . r
p p p . . p p p
. . n . . . . .
. . . n p . . .
. . B . . . . .
. . . . . N . .
P P P P . P P P
R N B Q R . K .

Stockfish preferred a quiet bishop move to e7. In the game White chose h3 instead, I traded my bishop for the knight, and a move later my knight left d5, opening the bishop's diagonal from c4 to f7. White's queen took on f7 and it was checkmate on move 9.

The lesson of the week

All three moments have the same shape: my move attacked something (a queen, a queen, a "pinned" knight) and I never looked at what my opponent could do back. So this is the routine I took from week 1, and it's how I'd tell any beginner to analyze their games:

1. Open the evaluation graph and find the two or three biggest drops. Ignore the rest.
2. Set up the position before the drop and cover the engine's answer.
3. List the forcing moves for both sides: checks, captures, attacks on the queen.
4. Only then compare with the engine's move and say in one sentence what it does.

At the board the same routine shrinks to one question before any attacking move: what are their checks and captures after this? Game #4 and game #5 would have lasted a lot longer. I turned this routine into a full step-by-step guide, using game #4 and game #7 as examples, in how to analyze chess games as a beginner.

All games this week

Next up is week 2, when I switched to faster games and lost nine in a row. The full story starts on the journey page, and if you'd like your own games reviewed like this, here is how the AI chess analysis works.