What is the Elo Rating System?
The Elo rating system is a method for calculating the relative skill levels of players in zero-sum games such as chess, Go, football, and many online video games. Invented by Hungarian-American physics professor Arpad Elo, the system assigns each player a numerical rating — the higher the number, the stronger the player.
Unlike win/loss records, Elo ratings update dynamically after every game: beating a stronger opponent earns you more points than beating a weaker one, while your opponent loses exactly as many points as you gain.
Elo Rating Definition
An Elo rating is a single number representing a player's estimated strength. New club-level players typically start around 1000–1200, intermediate players around 1400–1800, and grandmaster-level chess players above 2500. The World Chess Champion Magnus Carlsen has held a peak rating above 2880.
How to Find Elo Rating Change?
After each game, a player's new rating is calculated as:
- Compute the expected score EA for the player based on the rating difference.
- Record the actual score SA: 1 for a win, 0.5 for a draw, 0 for a loss.
- Apply the formula: R' = R + K × (SA − EA)
The K-factor controls how much the rating can change in a single game. A higher K means faster rating changes (used for beginners), while a lower K makes ratings more stable (used for top players).
Statistical Basics of Elo Rating
The expected score formula is derived from the logistic distribution:
EA = 1 / (1 + 10(RB − RA) / 400)
This means:
- If both players have the same rating, each has a 50% chance of winning.
- A rating difference of 200 points gives the stronger player roughly a 76% win probability.
- A difference of 400 points gives approximately a 91% win probability.
A Detailed Analysis — The Meaning of Elo Coefficients
The K-factor is the most important coefficient in the Elo system. Different organizations use different K-values:
- FIDE (International Chess Federation):
- K = 40 — for new players with fewer than 30 rated games or a rating below 1000
- K = 20 — for players rated below 2400
- K = 10 — for players rated 2400 and above who have played at least 30 games
- USCF (United States Chess Federation) — American System:
- K = 32 — for players rated below 2100
- K = 24 — for players rated 2100–2399
- K = 16 — for players rated 2400 and above
Our calculator supports both FIDE (international / metric system) and USCF (American system) K-factor rules, as well as a custom K-factor for other rating organizations.
Elo Calculator in Practice — Elo Rating in a Chess Tournament
Suppose you have a rating of 1600 and you play three games in a tournament using FIDE rules (K = 20):
- Win against a 1500-rated opponent — expected win probability: 64%, rating change: +7.2
- Draw against a 1650-rated opponent — expected win probability: 43%, rating change: +2.9
- Loss against a 1750-rated opponent — expected win probability: 31%, rating change: −6.3
After three games, your new rating would be approximately 1604 — a modest gain despite a mixed result, because you performed better than statistically expected.
Other Rating Systems
The Elo system inspired many variants and successors:
- Glicko / Glicko-2 — adds a "rating deviation" to measure confidence in the rating. Used by many online chess platforms.
- TrueSkill — Microsoft's system for Xbox Live, designed for team games.
- Elo in football — the World Football Elo Ratings use K = 60 for World Cup matches.
- Online games — League of Legends (MMR), Counter-Strike, and many other games use Elo-based ranking.
FAQs
What does a negative Elo change mean?
It means you performed worse than statistically expected. Even a loss against a much stronger opponent will cost you fewer points than expected.
Can my rating go below zero?
In theory, yes — but most organizations set a floor of 100 or 1000 to prevent negative ratings.
How many games does it take to get an accurate Elo?
Ratings become reliable after about 30–50 games. Before that, ratings are provisional and can swing significantly.
Is there a difference between the FIDE and USCF systems?
Yes. The USCF (American system) uses higher K-factors (32/24/16) for all but the very top players compared to FIDE (20/10). This means USCF ratings tend to fluctuate more rapidly in response to results.