Elo rating is a single number that estimates a player's current strength from their match results. Originally built for chess, it works for tennis because the only input it needs is who beat whom. Every win and loss nudges a player's number up or down.
How it works
Every player carries a rating. The gap between two ratings maps directly to a win probability. After each match the winner gains points and the loser sheds the same amount, scaled by how surprising the result was: beating a far higher-rated opponent moves your rating a lot, while beating someone you were expected to beat moves it a little.
The win-probability formula is:
expected score = 1 / (1 + 10^((opponent_rating - player_rating) / 400))
Worked example. If Player A is rated 1900 and Player B is rated 1700, the 200-point gap implies A wins about 76% of the time. If B pulls the upset, B's rating jumps and A's falls; if A wins as expected, both barely move.
The K-factor controls how much a single result shifts a rating. A higher K reacts faster to recent form but is noisier; a lower K is more stable but slower to recognize a rising player.
Why Smashrs uses it
Elo updates continuously and reflects current level, where official ranking points lag months behind and double-count tournament prestige. We rate every surface separately (see Surface Elo) and feed those ratings into the model edge calculation. For how Elo diverges from the official lists, see our ATP rankings and the Elo vs ATP ranking writeup.