Chess improvement
Chess Elo Rating Explained: How It Works
The Elo rating system predicts the expected score of a game from the rating gap between two players, then moves each rating by K × (actual score − expected score). A 200-point favourite is expected to score about 76%. FIDE uses Elo; Chess.com and Lichess use the related Glicko systems, so ratings are only comparable within the same pool and time control.
Updated

Your preparation checklist
- Elo compares your actual result with the result your rating gap predicted.
- FIDE moves ratings with a K-factor of 40, 20 or 10 depending on experience, age and level.
- Compare ratings within the same platform and time control.
- Treat rating milestones as checkpoints, not universal skill definitions.
- Use recurring mistakes in your games to choose what to study next.
What does the number say about your chess ability?
You have just passed 1000 Elo. Or perhaps you have been stuck at 1500 for several months. Either way, the same question comes up: what does that number actually say about your chess ability?
Your rating helps you compare your results with those of other players. But it does not tell you exactly what your strengths are, where you struggle, or what you should study next.
To make your rating useful, you need to understand what it measures—and look at the games behind the number.
What is an Elo rating?
The Elo rating system was designed by the physicist Arpad Elo and adopted by FIDE in 1970. It gives each player a number, and the difference between two numbers predicts how a game between them should go.
After each game, ratings move according to the difference between the expected result and the actual result. Beating a higher-rated opponent earns more points than beating a lower-rated one, and losing to a stronger player costs fewer points.
Your rating is therefore an estimate of your results within a particular player pool. That estimate becomes more reliable as you play more games.
How Elo is calculated: the formula
Two steps turn a game result into a rating change. First, the expected score of player A against player B is E = 1 / (1 + 10^((Rb − Ra) / 400)), where a win counts 1, a draw 0.5 and a loss 0.
Then the new rating is R′ = R + K × (S − E), where S is the actual score and K controls how fast ratings move.
Every 400 points of difference multiplies the favourite’s expected odds by ten. A few rating gaps and their expected scores:
| 0 | 50% |
|---|---|
| 100 | 64% |
| 200 | 76% |
| 300 | 85% |
| 400 | 91% |
Worked example with K = 20: a 1600 player meets an 1800 player. The 1600 player’s expected score is 1 / (1 + 10^0.5) ≈ 0.24. A win gains 20 × (1 − 0.24) ≈ +15 points, a draw ≈ +5 and a loss ≈ −5. The 1800 player gains or loses the same amounts in the other direction.
K-factor: how fast a FIDE rating changes
FIDE uses three K-factors, so new and young players move faster than established ones:
| 40 | New players until they complete 30 rated games, and players under 18 rated below 2300 |
|---|---|
| 20 | Players whose rating stays below 2400 |
| 10 | Players who have reached a published rating of 2400 |
With K = 10, the same upset win from the example above would bring about +8 points instead of +15. This is why ratings near the top move slowly.
Why your Chess.com, Lichess, and FIDE ratings differ
FIDE uses Elo. Chess.com uses Glicko and Lichess uses Glicko-2, two refinements of Elo that also track how certain each rating is, so a new or inactive player’s rating moves faster. Players still commonly say “Elo” for all of them.
A rating only makes sense within the pool where it is calculated. Platforms have different player populations and rating settings. They also maintain separate ratings for time controls such as blitz and rapid.
You can therefore have several different ratings without any one of them being your definitive “true rating.”
Time controls also test different skills. In blitz, you need to recognize patterns quickly and make decisions under pressure. In slower games, you have more time to calculate and check your ideas.
To track progress, focus on the same platform and time control over a meaningful period, rather than comparing isolated numbers across different systems.
What is a good chess Elo rating?
“Good” depends on the pool, but FIDE titles give fixed reference points. Each title requires the rating below plus results (norms) for the higher titles:
| Grandmaster (GM) | 2500 |
|---|---|
| International Master (IM) | 2400 |
| FIDE Master (FM) | 2300 |
| Candidate Master (CM) | 2200 |
| Woman Grandmaster (WGM) | 2300 |
| Woman International Master (WIM) | 2200 |
| Woman FIDE Master (WFM) | 2100 |
| Woman Candidate Master (WCM) | 2000 |
On the October 2026 FIDE list, 2000 puts you in the top 12.6% of active players and 2200 in the top 3.8%. The highest FIDE rating ever recorded is Magnus Carlsen’s 2882, reached in 2014. To see where real players stand, look up any player’s FIDE rating or browse the top-rated players of each federation.
What do 1000, 1500, and 2000 Elo mean?
Assigning the same skills to everyone at a particular rating would be misleading. Two players rated 1500 can achieve similar results through very different strengths: one might calculate well, while the other understands positional play better.
The following checkpoints are prompts for reviewing your games, rather than universal definitions of each rating.
Around 1000: check threats and hanging pieces
A useful starting point is to look for avoidable material losses.
You have a plan, play a move to prepare it, and then discover that your opponent was threatening your knight.
Before each move, build the habit of checking:
- What threat did my opponent’s last move create?
- Is any of my material attacked or insufficiently defended?
- After my intended move, which reply would cause me trouble?
If these mistakes appear regularly in your games, this checking routine deserves a central place in your training.
Around 1500: connect calculation with a plan
As immediate mistakes become less frequent, another difficulty may emerge: finding a direction when there is no obvious tactic.
You develop your pieces comfortably, then hesitate between several reasonable moves. Your decisions follow one another without working toward a shared objective.
Start with concrete questions:
- Which of my pieces is least active?
- Which weakness can I target?
- Which pawn break would improve my position?
Then check your plan tactically. An attractive strategic idea still needs to survive your opponent’s immediate responses.
Around 2000: examine precision and consistency
For an experienced player, useful training priorities may lie in more subtle decisions: converting an advantage, defending a difficult position, or managing the clock.
After a game, identify a specific decision you could improve:
- Did I simplify into an endgame that was actually favorable?
- Did I allow unnecessary counterplay while pursuing my attack?
- Did I spend too much time before the critical moment?
These questions can help players at any rating. Your own games determine which ones deserve priority.
Three types of positions that reveal what to work on
A single position cannot reliably establish someone’s rating. It can, however, reveal a skill worth developing.
An undefended piece and a possible fork: the task is to identify targets and check forcing moves—checks, captures, and threats.
A quiet position with no immediate material gain: the task is to compare piece activity, identify a weakness, and propose a practical plan.
An endgame with an extra pawn: the task is to work out how to make progress, which pieces to exchange, and what defensive resources to anticipate.
The useful question is why you found or missed the idea. Was the difficulty caused by missing knowledge, faulty calculation, attention, or time pressure?
How to choose your next training goal
Start by reviewing about ten games played at your usual time control. Before opening an engine, note where you hesitated and what you intended to do.
Then look for recurring mistakes:
| Losing pieces in one move | Threat checks and straightforward tactical exercises |
|---|---|
| Miscalculating variations | Short calculation exercises that include the opponent’s strongest reply |
| Struggling to find a plan | Annotated games and exercises on piece activity |
| Letting advantages slip | Conversion practice and endgames from your own games |
| Rushing decisions in time trouble | Clock management and reviewing long thinking periods |
Choose one priority and work on it for a few weeks. Then return to your recent games: does the mistake happen less often? Are your decisions clearer?
Your rating shows results; your games show what to study
Reaching a rating milestone is a motivating goal. To make progress toward it, translate that goal into a specific skill: spotting threats more reliably, calculating more accurately, or converting more favorable positions.
The most useful question becomes: “Which recurring mistake can I start correcting today?”
Frequently asked questions
How is chess Elo calculated?
Your expected score is 1 / (1 + 10^((opponent rating − your rating) / 400)). After the game your rating changes by K × (actual score − expected score), where a win is 1, a draw 0.5 and a loss 0. FIDE uses K = 40, 20 or 10.
What is a good Elo rating in chess?
It depends on the pool. FIDE titles are useful landmarks: 2000 for Woman Candidate Master, 2200 for Candidate Master, 2300 for FIDE Master, 2400 for International Master and 2500 for Grandmaster.
What is the highest Elo rating ever?
Magnus Carlsen reached a FIDE rating of 2882 in 2014, the highest ever recorded.
Is 1500 Elo the same on every platform?
No. Ratings belong to a particular player pool and time control. Compare your progress within the same platform and cadence rather than treating different ratings as interchangeable.
Can one position reveal my chess rating?
A single position cannot reliably establish your rating. It can help identify a skill to work on, such as calculation, planning or endgame technique.
What should I study to improve my rating?
Review about ten recent games at your usual time control, identify a recurring mistake and choose one training priority. Revisit your games after a few weeks to assess whether that mistake appears less often.
Sources & further reading
Rating-system facts are linked to the source above. Rating checkpoints and training suggestions are editorial guidance, not measured skill thresholds or promises of rating gains.






