ELO is a rating system that adjusts after every match based on what was expected to happen. Beat someone rated well above you and your rating jumps; beat someone below you and it barely moves. Lose to a stronger player and you lose almost nothing. Over enough matches, the number converges on something no win-loss record can give you: an honest measure of how good you actually are, relative to the people you play.
It was invented for chess by Arpad Elo in the 1960s and has since spread to basically every competitive game — because it solves the exact problem every tennis and padel group has: win counts lie.
Why win-loss records are propaganda
Your friend with the 9–2 record who only ever plays the group's two weakest players is not the best player in the group. He's the best scheduler. A raw win percentage rewards cherry-picking opponents; ELO punishes it. Farm easy wins and each one is worth a point or two. Take on the group's top seed and lose narrowly, and you've lost nearly nothing — beat them once and you leapfrog three people.
That single property changes group behavior. People start playing up. Rematches mean something. The ladder becomes real.
How the math works (without the math)
Every player starts at a baseline rating. Before each match, the gap between the two sides' ratings implies a win probability — a large gap means the favorite is expected to win comfortably. After the match, ratings move in proportion to the surprise: expected results move numbers a little, upsets move them a lot. Some systems (Rackt included) also weigh the score margin, so a 6-0 6-1 statement win moves the table more than a 7-6 7-6 escape.
In doubles, the pair's combined strength faces the opposing pair's, and each player's individual rating absorbs the result — so your rating is yours, even when your partner rotates every Americano round.
Getting an ELO for your actual group
Booking platforms rate you against their whole user base on a universal scale, which is useful for finding strangers to play. But your group doesn't care where you stand among two million users. It cares whether you've finally passed your rival.
That's what Rackt does: a private ELO ladder built only from your group's confirmed matches. Log the result after the handshake, your opponent confirms it (no mystery edits), and the table moves. Streaks, head-to-heads, form over the last seven matches — all computed from real results the whole group signed off on. Download Rackt free and give the group chat a number to argue about.
What a rating actually settles
Whether you're better than your rival: settled by the head-to-head and the ratings. Whether you're improving: settled by the trend line. Whether last night's tournament run was a hot streak or a new level: give it three more sessions and the table will tell you. The one thing ELO can't settle is who has the best excuses — those remain unranked.
FAQ
How many matches until my ELO is accurate?
Roughly 10–15 results gets the number close; it keeps sharpening after that.
Does ELO work for doubles with rotating partners?
Yes — that's its quiet superpower. Ratings are individual, so formats like Americano feed the ladder perfectly.
Is a padel ELO comparable to a tennis ELO?
No — a rating only means something within the pool that produced it. Rackt keeps separate tables per sport for exactly that reason.
Can my rating go down if I stop playing?
In Rackt, ratings move only on match results. Hiding, however, is visible to everyone in your match count.
Rackt is the match tracker for tennis and padel crews — private ELO rankings, confirmed scores, rivalries and tournament nights. Free on iOS and Android.