How Elo ratings work
Understand expected results, rating changes and K factors in a Rankerelo leaderboard.
A rating estimates relative results
Elo compares two players’ current ratings to estimate an expected result. Winning against a higher-rated opponent usually gains more points than winning against a lower-rated opponent.
The calculation
For player A, the expected score is 1 / (1 + 10^((rating B − rating A) / 400)). The new rating is the old rating plus K × (actual result − expected result), rounded to a whole number. A win scores 1, a draw 0.5 and a loss 0.
A simple example
With two players rated 1000 and K = 32, each is expected to score 0.5. The winner gains 16 points and the loser loses 16. A draw between equally rated players leaves both ratings unchanged.
What the K factor changes
A higher K factor makes ratings react more quickly; a lower one makes them more stable. A board can choose its K factor. Supported result corrections use the K factor recorded when each match was applied.
What ratings cannot tell you
Ratings reflect the recorded results within a particular group. They are not an official qualification, and ratings from different boards need not be directly comparable.