How Football Elo Ratings Work

Elo ratings, invented for chess by Arpad Elo, are one of the most reliable ways to compare football teams that rarely meet. Here is how ours work, in plain English.

The core idea

Every team has a rating. Before each match, the difference between the two ratings gives an expected result: a team rated 400 points higher is expected to score about 91% of the available “points” (a win counts as 1, a draw as 0.5). After the match, each team gains or loses points in proportion to how much the real result differed from the expectation. Beating a much stronger team earns a lot; beating a much weaker team earns little.

The formula

New rating = old rating + K × G × (W − We), where:

W is the result (1 win, 0.5 draw, 0 defeat; shoot-outs count as draws);
We is the expected result, 1 / (10^−d/400 + 1), with d the rating difference plus 100 points for the home team;
K reflects the importance of the match: 60 for World Cup finals, 50 for continental finals such as AFCON, 40 for qualifiers and Nations League games, 30 for other tournaments and 20 for friendlies;
G rewards the margin: 1 for a one-goal win, 1.5 for two goals, and (11 + N) / 8 for wins by three or more.

Why Elo works well for Africa

Many African teams play few matches against the world's top sides, which makes simple win-loss tables misleading. Elo links every team through the network of all results since 1872, so a strong run in AFCON qualifying against well-rated opponents is properly rewarded. Since 2018 FIFA's own world ranking also uses an Elo-style formula, though with different weights.

Our implementation

We recompute ratings from scratch after every data update. Teams start at 1500 points and their first 15 matches count double, so newer nations such as South Sudan quickly reach a realistic level. Teams that have not played for four years are left out of the current ranking table but keep their history.

See the live African Elo rankings, or open the “Elo rating” tab on any national team page for its full history.