Rank My PadelEngine v1.0.0 · no login needed

How your rating is worked out.

Every player has a number from 0.00 to 7.00 and a reliability from 0 to 100%. Nothing here is simplified for public consumption — this is what the code does, and the widget below runs the same functions that rate real matches.

The formula, in five sentences

Average each pair’s two ratings. That gives two team numbers, and the gap between them is all that matters — not who is on which side of it.
The gap predicts what share of games each pair should win. Half a level is worth about 9.5 percentage points.
Compare that prediction to the share you actually won. The difference is the surprise. Who won is never an input — only the margin against expectation.
The surprise becomes a rating for that match: your team’s average, plus 3 × the surprise. One match, one number, arguing for a level.
Your rating is the weighted average of all of them, plus your placement, recomputed from scratch every time. Nothing is added to a running total, which is why nobody — including us — can reach in and adjust it.

That fifth sentence is why an established player moves in hundredths and a new one moves in tenths: with forty matches behind it, one more barely shifts the average. There is no separate rule that slows you down, and no cap on the rating change — a genuinely surprising result is allowed to say so. The one bound anywhere near this is on the SCORE: a whitewash is read as 98% of games rather than 100%, so a single perfect scoreline cannot imply an infinite gap in ability.

Change the scoreline and watch it move

The same code path that rates real matches, running in your browser. Nothing here is a mock-up of the maths.

Your pair

average 4.20

Theirs

average 4.50

Games won

–

Expected44%

You took56%

Gap between the pairs-0.30

Expected share44.3%

Actual share56.3%

Surprise+12.0 points

What this match argues your level is4.56

That last number is not your new rating. Your rating is the weighted average of every match you hold and your placement, so this one pulls it that way by however much weight it carries — format, verification, how close the match was, and how recently you played it.

Why you can lose and go up

Expected to take 26% of games and you took 43%? You were better than the prediction, so the number rises — even though the scoreboard says you lost.

Why you can win and go down

Beat a much weaker pair 6-4 when the maths expected 6-1, and you have underperformed a prediction you were always going to beat. The number edges down.

What we deliberately don’t do

No decay for taking a month off. Your rating does not fall because you stopped playing. What falls is our confidence in it, because we weight recent evidence more heavily — the number stays exactly where you left it.
No cap on a single match. If a result is surprising, it is allowed to be. Nothing clips the change.
No points for turning up. Playing more does not raise your rating; playing better does. More matches raise your reliability.
No hidden adjustments. The rating is a function of the match log. Rebuild the log and you get the same number, every time.
No self-assessment after the first day. The questionnaire gets you a starting point and then washes out.
No separate ladder. Bands are labels for conversation, not tiers to climb.

The scale

8 bands, with boundaries following Playtomic’s published chart so the two numbers are directly comparable.

Initiation0.00 – 1.00
Beginner1.00 – 1.50
Improver1.50 – 2.50
Intermediate2.50 – 3.50
Upper Intermediate3.50 – 4.50
Advanced4.50 – 5.40
Competitor5.40 – 6.00
Elite6.00 – 7.00