Turn those games into a training plan

Import from Lichess or Chess.com Β· no signup needed

← Back to Learning
Improve by Rating🌱Beginnerβ€’ 9 min read

How to Get Better at Chess: What Really Changes From 400 to 2100 Elo

We analysed more than 30,000 Chess.com player-games with Stockfish to see what separates each rating level. Surprise: stronger players blunder almost as often. What changes is when, and how.

The short version. Below 1000, you lose games to pieces left hanging in the first ten moves. Between 1000 and 1500, you lose them to tactics in the middlegame. Above that, mostly to positional mistakes and endgames. Train the mistake that matches your level, not the one in the generic advice.

Where These Numbers Come From

We looked at the games imported into ChessMentor AI from Chess.com: the games of our users and of their opponents. Every move was analysed with Stockfish and classified by how much it changed the player's winning chances:

  • Blunder: the move drops the winning chances by 22 percentage points or more.
  • Mistake: a drop of 13 to 22 points.
  • Inaccuracy: a drop of 6 to 13 points.

Each player-game is counted once per side, at that player's Chess.com rating. In total: more than 30,000 player-games, over 1 million analysed moves, from around 12,000 different players.

Below 1200, most games are rapid; higher up, blitz becomes more common. Keep that in mind when comparing levels.

The Big Table

Rating Blunders per game Mistakes per game Games with at least one blunder Move of the first blunder (median) Blunders made in the opening Blunders that hang a piece
Under 600 1.97 1.34 77% 8 58% 16%
600–800 1.93 1.52 75% 11 48% 13%
800–1000 1.63 1.40 73% 11 43% 14%
1000–1200 1.53 1.40 70% 11 47% 7%
1200–1500 1.56 1.53 71% 16 29% 8%
1500–1800 1.50 2.03 72% 17 25% 7%
1800–2100 1.38 2.35 68% 18 22% 9%
2100+ 1.33 1.79 65% 20 17% 8%

"Hangs a piece" means the opponent's best reply simply takes an undefended knight, bishop, rook or queen.

Five Things the Data Says

1. Stronger players still blunder

A player rated under 600 makes about 2 blunders per game. A player above 2100 still makes 1.3. Two games out of three contain at least one blunder at every level.

So "stop blundering" is not a realistic goal. The real goal is to blunder later, in less obvious ways, and less often in decisive moments.

2. The first blunder comes later and later

The median first blunder arrives at move 8 below 600, move 11 between 600 and 1200, move 16 around 1200–1500, and move 20 above 2100.

Every few moves you survive without a blunder give you more time to outplay your opponent. That is what rating points are made of.

3. Beginners lose games in the opening

Below 600, 58% of blunders happen in the opening, and 39% in the first ten moves alone. By 1200–1500 it is 29%, and above 2100 only 17%.

This is not about opening theory. It is about basic safety in the first moves: developing pieces to safe squares and noticing early attacks.

4. Hanging pieces disappear around 1000

Below 1000, about one blunder in seven simply leaves a piece where it can be taken for free. From 1000 upward, this drops to about one in thirteen, and stays there.

That is the clearest dividing line in the data: if you stop hanging pieces, you will very likely pass 1000.

5. Most hung pieces are not the piece that just moved

When a piece gets taken for free, two times out of three it is not the piece the player just moved. Either it was already under attack and the player ignored the threat, or the move took away its last defender.

The fix is a habit, not knowledge: before every move, ask "what does my opponent's last move attack?" and "what does the piece I'm about to move currently protect?"

What to Train at Your Level

Pick your rating range for a detailed guide with real examples from games at that level:

The Fastest Way to Know Your Own Pattern

Averages describe thousands of players. Your games have their own pattern: maybe you hang pieces after move 20, or blow won endgames, or fall for the same fork again and again.

That's what ChessMentor AI measures. Import your Chess.com or Lichess games, and it shows the mistakes you repeat most, with a style radar and one fully coached game for free.

William Mbotta Elimbi

William Mbotta Elimbi

Top 5% Lichess Β· Top 0.5% Chess.com

Elite chess player and founder of ChessMentor AI. Combining Stockfish-powered analysis with LLM coaching to help players improve through personalized annotations and training.

β™Ÿ

Put it into practice β€” Analyse your games with the AI coach

Import your games β€” style radar and 1 coached game free.

Analyse my games β†’

πŸ“¬ Get New Chess Articles

Subscribe to receive my latest chess guides and opening strategies directly in your inbox!

No spam. Unsubscribe anytime. Your privacy is our priority.

πŸ“’ Share this article

Free Β· No signup neededπŸ”’ Your games stay private

Turn those games into a training plan

Import Lichess or Chess.com. Get a style radar, puzzles from your own mistakes, and 1 coached game β€” free.

!!Brilliant
βœ“Good
?Mistake
??Blunder
Accuracy
Import my games β€” free β†’