How good is a 2900 chess player?
Rates are per 100 live moves: moves played while the game is not already decided.
What the terms meanShow less
- Centipawn.
- A hundredth of a pawn, the unit a chess engine scores a position in. Giving up a pawn for nothing costs about 100.
- Centipawn loss.
- How much worse a move is than the best move in the position. The average (ACPL) counts every move a player makes.
- Blunder, mistake, inaccuracy.
- A move that loses at least 150 centipawns is a blunder, at least 70 a mistake and at least 40 an inaccuracy. Only live moves are counted.
- Only move.
- A position where one move is clearly better than every other. “Finds the only move” is how often a player plays it.
- Time trouble.
- Less than a tenth of the starting time left on the clock. “Games in time trouble” is the share of games in which a player's clock got that low on any move, however the game stood.
- Winning, better, balanced, worse.
- How the engine scores the position for the player about to move. Balanced is within half a pawn either way, better is up to a pawn and a half ahead, winning is from there up to three pawns ahead, and worse is up to a pawn and a half behind.
- +1, +2, +4.
- A lead of at least one, two or four pawns by the engine's score. A lost position is one that has fallen two pawns behind or worse.
- Percentile.
- The share of a site's players who are rated lower. Blitz and rapid players are ranked separately.
How strong is 2900 Elo?
Over99.99%of Chess.com blitz players are rated below 2900
Fewer than 0.01% are rated 2900 or higher: about 820 players.
5.0blunders per 100 live moves
8.5inaccuracies per 100 live moves
87%of only moves found
38%of games reach time trouble
Chess.com blitz 3+0Chess.com blitz players at each rating
Chess.com blitz 3+0Chess.com blitz players at each rating
Chess.com blitz 3+0Chess.com blitz players at each rating
Chess.com blitz 3+0Chess.com blitz players at each rating
What a 2900 actually does
| Measure | 2800 | 2900 | 3000 |
|---|---|---|---|
| Blunders | 5.3 | 5.0 | 4.5 |
| Mistakes | 8.3 | 7.8 | 7.4 |
| Inaccuracies | 9.1 | 8.5 | 8.5 |
| Average centipawn loss | 36.4 | 34.3 | 32.5 |
| Plays the engine's top move | 44% | 45% | 45% |
| Finds the only moveshare of positions with one real answer | 86% | 87% | 88% |
| Game length | 45.5 | 45.5 | 47.2 |
| Live moves per gamemoves played while the game is not already decided | 31.1 | 31.6 | 33.6 |
| How they use the clock | |||
| Games in time troubleunder a tenth of the clock | 40% | 38% | 39% |
| Seconds per move | 3.5 | 3.5 | 3.4 |
| Instant movesplayed in two seconds or less | 59% | 60% | 61% |
| Clock left at move 8share of the starting clock | 92% | 92% | 93% |
| Clock left at move 15share of the starting clock | 77% | 77% | 78% |
| Clock left at move 30in games that get there | 37% | 38% | 39% |
| Clock left at move 50in games that get there | 14% | 14% | 15% |
| Moves played inside a scrambleown clock under a tenth | 8.5% | 8.2% | 8.9% |
| What the game looks like | |||
| Games that end decisivelywon or lost, not drawn | 92% | 91% | 90% |
| Reaches an endgame | 45% | 46% | 47% |
| Capturesper 100 own moves | 21.7 | 21.7 | 21.2 |
| Checksper 100 own moves | 6.9 | 6.7 | 6.9 |
How others play against a 2900
| Measure | 2800vs 2900 | 2900vs 2900 | 3000vs 2900 |
|---|---|---|---|
| Score against them | 39.7% | 50.0% | 59.2% |
| Score after reaching +1 | 66% | 73% | 78% |
| Score after reaching +2 | 75% | 80% | 85% |
| Ever loses from +4 | 10% | 9% | 5% |
| Score from lost positionsgames where their eval fell to minus two | 20% | 23% | 29% |
Each column is how players of that rating do in their games against a 2900.
Where a 2900 blunders
| Measure | 2800 | 2900 | 3000 |
|---|---|---|---|
| When in the game | |||
| Opening | 0.8 | 0.7 | 0.5 |
| Middlegame | 6.3 | 5.8 | 5.0 |
| Endgame | 9.2 | 8.9 | 8.4 |
| In what kind of position | |||
| Winning | 8.8 | 9.0 | 7.7 |
| Better | 3.2 | 3.1 | 2.4 |
| Balanced | 2.8 | 2.7 | 2.3 |
| Worse | 4.9 | 4.7 | 4.3 |
| With how much clock left | |||
| Over half | 2.7 | 2.5 | 2.0 |
| 25 to 50% | 7.3 | 6.4 | 5.6 |
| 10 to 25% | 9.5 | 9.4 | 8.2 |
| Under 10% | 14.2 | 14.1 | 12.9 |
| With which piece | |||
| Queen | 7.4 | 6.7 | 6.0 |
| Rook | 6.8 | 6.2 | 5.7 |
| Knight | 4.6 | 4.6 | 4.0 |
| Bishop | 4.5 | 4.1 | 3.6 |
| Pawn | 4.2 | 4.0 | 3.4 |
| King | 5.5 | 5.5 | 5.0 |
The four rows under “In what kind of position” leave out positions that are already won or already lost. Throwing away a won game still counts in the overall blunder rate, which is why that rate can be higher than all four.
When a 2900 blunders most
Blunders per 100 live moves
When in the game
Most in the endgame: 13 times as often as in the opening.
- 2900, Opening: 0.7
- 2900, Middlegame: 5.8
- 2900, Endgame: 8.9
In what kind of position
Most when winning: 3.3 times as often as in a balanced game.
- 2900, Winning: 9.0
- 2900, Better: 3.1
- 2900, Balanced: 2.7
- 2900, Worse: 4.7
With how much clock left
Most with under 10% of the clock: 5.6 times as often as with over half the clock.
- 2900, Over half: 2.5
- 2900, 25 to 50%: 6.4
- 2900, 10 to 25%: 9.4
- 2900, Under 10%: 14.1
With which piece
Most on queen moves: 68% more often than on pawn moves.
- 2900, Queen: 6.7
- 2900, Rook: 6.2
- 2900, Knight: 4.6
- 2900, Bishop: 4.1
- 2900, Pawn: 4.0
- 2900, King: 5.5
How much 2900-rated players differ
| Measure | 5% | 25% | 75% | 95% |
|---|---|---|---|---|
| Blunders | 4.1 | 4.7 | 5.6 | 6.3 |
| Mistakes | 6.8 | 7.4 | 8.3 | 9.1 |
| Inaccuracies | 7.4 | 8.0 | 8.9 | 9.6 |
| Finds the only move | 83% | 86% | 89% | 91% |
| Seconds per move | 3.0s | 3.2s | 3.6s | 4.0s |
| Games in time trouble | 20% | 28% | 46% | 57% |
Each column is a share of 2900-rated players: a quarter are at or under the 25% figure, three quarters at or under the 75% one.
The move quality range a 2900 plays in
2750 to 3000is how half of 2900-rated players play, going by how often they blunder
Nine in ten: 2600 to 3150.
Of 100 players rated 2900, how many blunder
- like a 2800 or lower33
- like a 290048
- like a 3000 or higher19
What separates a 2900 from a 3200+
| Skill, and what trains it | 2900 | 3200+ | Gap |
|---|---|---|---|
| Opening blundersthe first ten moves of your openings | 0.7 | 0.3 | 54% lower |
| Middlegame blundersa check-captures-threats scan every move | 5.8 | 4.0 | 31% lower |
| Finds the forkknight and queen forks, both directions | 65% | 79% | +14 pts |
| Endgame blundersrook endings and king activity | 8.9 | 7.3 | 19% lower |
| Sees the deeper wincalculating past the third move of a line | 61% | 70% | +9 pts |
| Composure after an errorresetting after you drop something | 3.1% | 2.7% | 13% lower |
| Trades down when aheadconverting an advantage by liquidating | 71% | 77% | +6 pts |
| Finds the defensive savedefending worse positions instead of collapsing | 73% | 76% | +3 pts |
| Finds the quiet best movepositions where nothing is forcing | 77% | 80% | +3 pts |
| Takes the free materialspotting hanging pieces before anything else | 95% | 96% | +1 pt |
What a 2900 plays
As White
- 1.e4 King's Pawn 47.6%, scores 49.3%
- 1.d4 Queen's Pawn 30.8%, scores 51.5%
- 1.Nf3 Réti / Zukertort 13.1%, scores 54.3%
- every other move 9%
Against 1.e4
- 1...c5 Sicilian Defense 36.5%, scores 47.2%
- 1...e5 Open Game 23.1%, scores 44.2%
- 1...c6 Caro-Kann Defense 13.5%, scores 49.1%
- every other move 27%
Against 1.d4
- 1...Nf6 Indian Defense 57.9%, scores 42.6%
- 1...d5 Closed Game 24.6%, scores 41.6%
- 1...e6, usually a French or Queen's Gambit 5.0%, scores 40.1%
- every other move 13%
2900 on each site and time control
| Played on | Blunders | Mistakes | Inaccs | ACPL | Only move |
|---|---|---|---|---|---|
| Chess.com blitz 3+0 | 5.0 | 7.8 | 8.5 | 34.3 | 87% |
| Chess.com blitz 5+0 | ladder ends at 2500 | ||||
| Chess.com rapid 10+0 | ladder ends at 2600+ | ||||
| Lichess blitz 3+02900 or higher | 3.9 | 6.7 | 7.8 | 29.5 | 91% |
| Lichess blitz 5+0 | ladder ends at 2500+ | ||||
| Lichess rapid 10+0 | ladder ends at 2600+ | ||||
A rating number only means something inside one pool. Full conversion tables
What punishes a 2900
Share of all blunders, by the idea that refuted them
Most often a piece left hanging: 17% of all blunders, against 17% at 2600.
- 2900, Piece left hanging: 17%
- 2900, Losing exchange: 11%
- 2900, Pawn or promotion: 10%
- 2900, Fork: 8%
- 2900, Quiet move: 8%
Frequently asked about 2900-rated players
Answered for Chess.com blitz 3+0, the choice at the top of the page.
Is 2900 a good chess rating?
Yes, 2900 is a very strong chess rating. Over 99.99% of Chess.com blitz players are rated below 2900, and the middle player there is rated about 485.
What percentile is 2900 Elo?
A 2900 rating is above the 99.99th percentile on Chess.com. Under 0.01% of Chess.com blitz players are rated 2900 or higher, which is about 820 players. On Lichess, 99.98% of blitz players are rated below 2900. Every rating side by side: the Chess.com rating guide.
How accurate is a 2900 rated chess player?
A 2900-rated player plays the engine's top choice on 45% of moves in Chess.com blitz 3+0 and loses 34.3 centipawns per move on average. Going up the ratings, 2900 is where the average loss first drops under 35 centipawns a move.
How many blunders does a 2900 make per game?
A 2900-rated player makes about 1.6 blunders per game in Chess.com blitz 3+0, counting a move that loses at least 150 centipawns while the game is not already decided. A 2600 makes about 1.7 and a 3200+ about 1.3. The count barely moves between them, although a 2900 blunders on 5.0 of every 100 live moves and a 2600 on 6.1: a 2900's games stay undecided for longer, so there are more moves to get wrong.
Do 2900 rated players blunder more in time trouble?
Yes, a great deal. With under 10% of the clock left, a 2900-rated player blunders 5.6 times as often as with more than half of it. A 2900 gets that low on time in 38% of games in Chess.com blitz 3+0. The effect is bigger higher up: 8.3 times as often for a 3200+.
What mistakes do 2900 rated players make?
17% of the blunders a 2900-rated player makes in Chess.com blitz 3+0 are a piece simply left hanging, the most common kind. That share is 17% at 2600 and 16% at 3200+. Blunders involving a pawn or promotion grow at this level: 10%, up from 8% at 2600.
When in the game do 2900 rated players blunder most?
A 2900-rated player blunders most in the endgame in Chess.com blitz 3+0: on 8.9 of every 100 live moves there, against 5.8 in the middlegame and 0.7 in the opening. The endgame is also played on the least clock, so much of that is time pressure rather than weak endgame play.
What openings do 2900 rated players play?
As White in Chess.com blitz 3+0, 47.6% of players rated around 2900 open with e4, the King's Pawn, and 30.8% open with d4, the Queen's Pawn. The King's Pawn is opened less as ratings rise: 48.3% around 2600, 41.9% around 3200+. Against e4, the most common reply is c5, the Sicilian Defense, played in 36.5% of games, ahead of e5, the Open Game, at 23.1%.
How does a 2900 do against a 3000?
A 2900 scores 40.8% against a 3000 in Chess.com blitz 3+0. Against a 2800 they score 60.3%.
2900 rating everywhere else
This page opens on Chess.com blitz 3+0, and the choice at the top changes every table on it. The same tables and answers for the other one are below.
Lichess blitz 3+0
What a 2900+ actually does in Lichess blitz 3+0
| Measure | 2800 | 2900+ |
|---|---|---|
| Blunders | 4.4 | 3.9 |
| Mistakes | 7.1 | 6.7 |
| Inaccuracies | 8.4 | 7.8 |
| Average centipawn loss | 32.0 | 29.5 |
| Plays the engine's top move | 47% | 47% |
| Finds the only moveshare of positions with one real answer | 89% | 91% |
| Game length | 45.2 | 46.8 |
| Live moves per gamemoves played while the game is not already decided | 32.3 | 33.9 |
| How they use the clock | ||
| Games in time troubleunder a tenth of the clock | 33% | 33% |
| Seconds per move | 3.4 | 3.3 |
| Instant movesplayed in two seconds or less | 65% | 66% |
| Clock left at move 8share of the starting clock | 93% | 93% |
| Clock left at move 15share of the starting clock | 78% | 77% |
| Clock left at move 30in games that get there | 39% | 39% |
| Clock left at move 50in games that get there | 14% | 15% |
| Moves played inside a scrambleown clock under a tenth | 7.5% | 8.8% |
| What the game looks like | ||
| Games that end decisivelywon or lost, not drawn | 87% | 84% |
| Reaches an endgame | 44% | 47% |
| Capturesper 100 own moves | 21.8 | 21.6 |
| Checksper 100 own moves | 7.0 | 6.9 |
How others play against a 2900+ in Lichess blitz 3+0
| Result | 2800 | 2900+ |
|---|---|---|
| How the game goes | ||
| Score against them | 35.3% | 50.0% |
| Converting and defending | ||
| Score after reaching +1 | 62% | 76% |
| Score after reaching +2 | 73% | 83% |
| Ever loses from +4 | 13% | 4% |
| Score from lost positionsgames where their eval fell to minus two | 12% | 32% |
Each column is how players of that rating do in their games against a 2900+. The 2900+ column is everyone rated 2900 or above.
Where a 2900+ blunders in Lichess blitz 3+0
| Measure | 2800 | 2900+ |
|---|---|---|
| When in the game | ||
| Opening | 0.6 | 0.4 |
| Middlegame | 5.0 | 4.4 |
| Endgame | 8.2 | 7.1 |
| In what kind of position | ||
| Winning | 7.6 | 7.0 |
| Better | 2.6 | 1.9 |
| Balanced | 2.3 | 2.0 |
| Worse | 4.3 | 3.1 |
| With how much clock left | ||
| Over half | 2.1 | 1.7 |
| 25 to 50% | 5.8 | 5.0 |
| 10 to 25% | 7.9 | 7.1 |
| Under 10% | 12.8 | 11.2 |
| With which piece | ||
| Queen | 6.2 | 5.4 |
| Rook | 5.7 | 5.1 |
| Knight | 3.8 | 3.2 |
| Bishop | 3.8 | 3.1 |
| Pawn | 3.3 | 3.0 |
| King | 4.9 | 4.6 |
The four rows under “In what kind of position” leave out positions that are already won or already lost. Throwing away a won game still counts in the overall blunder rate, which is why that rate can be higher than all four.
How much players rated 2900 or higher differ in Lichess blitz 3+0
| Measure | 5% | 25% | 75% | 95% |
|---|---|---|---|---|
| Blunders | 3.0 | 3.5 | 4.3 | 4.8 |
| Mistakes | 5.6 | 6.2 | 7.1 | 7.9 |
| Inaccuracies | 6.3 | 7.1 | 8.2 | 9.0 |
| Finds the only move | 87% | 90% | 93% | 95% |
| Seconds per move | 2.7s | 3.0s | 3.5s | 4.0s |
| Games in time trouble | 13% | 22% | 42% | 57% |
Each column is a share of players rated 2900 or higher: a quarter are at or under the 25% figure, three quarters at or under the 75% one.
What separates a 2600 from a 2900+ in Lichess blitz 3+0
| Skill, and what trains it | 2600 | 2900+ | Gap |
|---|---|---|---|
| Opening blundersthe first ten moves of your openings | 0.9 | 0.4 | 59% lower |
| Middlegame blundersa check-captures-threats scan every move | 6.6 | 4.4 | 33% lower |
| Endgame blundersrook endings and king activity | 9.5 | 7.1 | 26% lower |
| Composure after an errorresetting after you drop something | 3.3% | 2.8% | 16% lower |
| Finds the defensive savedefending worse positions instead of collapsing | 69% | 79% | +10 pts |
| Trades down when aheadconverting an advantage by liquidating | 67% | 76% | +9 pts |
| Finds the quiet best movepositions where nothing is forcing | 76% | 83% | +7 pts |
| Sees the deeper wincalculating past the third move of a line | 62% | 71% | +9 pts |
| Takes the free materialspotting hanging pieces before anything else | 94% | 96% | +2 pts |
| Finds the forkknight and queen forks, both directions | 74% | 75% | +1 pt |
What a 2900+ plays in Lichess blitz 3+0
As White
| First move | Share | Score |
|---|---|---|
| 1.e4King's Pawn | 50.9% | 68.5% |
| 1.d4Queen's Pawn | 30.3% | 76.3% |
| 1.Nf3Réti / Zukertort | 9.2% | 72.6% |
| Every other move | 10% | n/a |
Against 1.e4
| Reply | Share | Score |
|---|---|---|
| 1...c5Sicilian Defense | 40.2% | 64.6% |
| 1...e5Open Game | 29.9% | 64.6% |
| 1...c6Caro-Kann Defense | 13.6% | 68.1% |
| Every other move | 16% | n/a |
Against 1.d4
| Reply | Share | Score |
|---|---|---|
| 1...Nf6Indian Defense | 62.6% | 70.2% |
| 1...d5Closed Game | 23.6% | 75.6% |
| 1...f5Dutch Defense | 3.9% | 57.1% |
| Every other move | 10% | n/a |
What punishes a 2900+ in Lichess blitz 3+0
| The idea that refuted the blunder | Share of blunders |
|---|---|
| Piece left hanging | 18% |
| Quiet move | 10% |
| Pawn or promotion | 9% |
| Losing exchange | 8% |
| Discovered attack | 7% |
Questions about a 2900+ in Lichess blitz 3+0
- Lichess blitz 3+0: Is 2900+ a good chess rating?
- Yes, 2900+ is a very strong chess rating. 99.98% of Lichess blitz players are rated below 2900, and the middle player there is rated about 1447.
- Lichess blitz 3+0: What percentile is 2900+ Elo?
- A 2900+ rating is above the 99th percentile on Lichess. It is in the top 0.02% of Lichess blitz players, which is about 240 players. On Chess.com, over 99.99% of blitz players are rated below 2900. Every rating side by side: the Lichess rating guide.
- Lichess blitz 3+0: How accurate is a 2900+ rated chess player?
- A player rated 2900 or higher plays the engine's top choice on 47% of moves in Lichess blitz 3+0 and loses 29.5 centipawns per move on average. Going up the ratings, 2900+ is where the average loss first drops under 30 centipawns a move.
- Lichess blitz 3+0: How many blunders does a 2900+ make per game?
- A player rated 2900 or higher makes about 1.3 blunders per game in Lichess blitz 3+0, counting a move that loses at least 150 centipawns while the game is not already decided. A 2600 makes about 1.6.
- Lichess blitz 3+0: Do 2900+ rated players blunder more in time trouble?
- Yes, a great deal. With under 10% of the clock left, a player rated 2900 or higher blunders 6.6 times as often as with more than half of it. A 2900+ gets that low on time in 33% of games in Lichess blitz 3+0. The effect is smaller lower down: 5.0 times as often for a 2600.
- Lichess blitz 3+0: What mistakes do 2900+ rated players make?
- 18% of the blunders a player rated 2900 or higher makes in Lichess blitz 3+0 are a piece simply left hanging, the most common kind. That share is 18% at 2600. Blunders involving a quiet move grow at this level: 10%, up from 8% at 2600.
- Lichess blitz 3+0: When in the game do 2900+ rated players blunder most?
- A player rated 2900 or higher blunders most in the endgame in Lichess blitz 3+0: on 7.1 of every 100 live moves there, against 4.4 in the middlegame and 0.4 in the opening. The endgame is also played on the least clock, so much of that is time pressure rather than weak endgame play. It takes until 2900 for the rate in the middlegame to drop under 5 in every 100 live moves.
- Lichess blitz 3+0: What openings do 2900+ rated players play?
- As White in Lichess blitz 3+0, 50.9% of players rated 2900 or higher open with e4, the King's Pawn, and 30.3% open with d4, the Queen's Pawn. Around 2600, e4 is opened by 45.6%. Against e4, the most common reply is c5, the Sicilian Defense, played in 40.2% of games, ahead of e5, the Open Game, at 29.9%.
- Lichess blitz 3+0: How does a 2900+ do against a 2800?
- A 2900+ scores 64.7% against a 2800 in Lichess blitz 3+0.