Elo calculation
Elo ranking built from real scorelines.
Every rated match updates ratings using the full game count, a weak link team strength model, and the classic Elo expected score curve. This page mirrors the exact math used in BallBuddy.
Games not sets
Weak link team strength
Weighted by match type
Current constants
These values come straight from the rating engine.
| Constant | Value |
|---|---|
| Default rating | 1200 |
| K factor | 32 |
| Weak link alpha | 0.45 |
| Friendly weight | 1 |
| Tournament weight | 1.2 |
Custom rating weight values are accepted when provided and are clamped to a minimum of 1 if a non-positive number is supplied.
The exact Elo math, step by step.
BallBuddy rates doubles by converting each team into a single strength value, then applying the Elo expected score curve. The rating change is proportional to the difference between the actual scoreline and the expected score.
Output applied to each player on a team
- Team 1 players
- + delta
- Team 2 players
- − delta
Formula block
All variables match the live rating engine.
teamStrength = min(r1, r2) + alpha * (avg(r1, r2) - min(r1, r2))
expectedTeam1 = 1 / (1 + 10 ^ ((strength2 - strength1) / 400))
actualTeam1 = team1Games / (team1Games + team2Games)
delta = weight * K * (actualTeam1 - expectedTeam1)
team1Players += delta
team2Players -= delta
Games matter. Blowouts move ratings more than close wins because actual score shifts.
Alpha controls balance. Lower alpha leans harder on the weaker player.
What goes into a rating change
Games decide the actual score
BallBuddy sums every set. A 6-4, 6-3 win becomes 12-7. The actual score is team1Games / totalGames.
Team strength uses a weak link model
We start from the lower rated player and pull toward the average using alpha to reflect team balance.
Expected score comes from Elo odds
We plug the two team strengths into the classic Elo curve to get a win probability.
K and match weight scale the change
K controls volatility. Match weight boosts tournaments or uses a custom value when set.
Translate Elo into a 1–10 skill label.
Elo remains a numeric rating, but you can summarize it as a simple padel level for quick communication. The scale is descriptive only and does not change the Elo calculation.
| Scale | Level | Elo range |
|---|---|---|
| 1 | Newcomer | 100–699 |
| 2 | Beginner | 700–999 |
| 3 | Class E | 1000–1199 |
| 4 | Class D | 1200–1399 |
| 5 | Class C | 1400–1599 |
| 6 | Class B | 1600–1799 |
| 7 | Class A | 1800–1999 |
| 8 | Pro | 2000–2199 |
| 9 | Master | 2200–2399 |
| 10 | Grandmaster | 2400–3000 |
Ranges follow chess-style distributions from 100 to 3000. Default Elo starts at 1200; adjust the ranges to fit your local community distribution.
Example from a real scoreline.
Team 1 (1300, 1100) defeats Team 2 (1200, 1200) by 6-4, 6-3. That is 12-7 in games. The numbers below use the default constants and the exact formula.
Computed strengths
Weak link model with alpha 0.45.
- Team 1 strength
- 1145
- Team 2 strength
- 1200
- Expected Team 1
- 0.422
- Actual Team 1
- 0.632
Rating delta
K 32, weight 1.
- Team 1 players
- +6.72
- Team 2 players
- −6.72
Tournament weight effect
- Delta with weight 1.2
- +8.07
Rating rules
Match must have recorded games
If total games are 0, the match is not rated.
Each player on a team gets the same delta
Both players on the winning team gain the same rating change.
Pair Elo uses the same formula
Pair ratings skip the weak link step and use the pair rating directly.
Ready to see your Elo move?
Create matches, record scores, and watch your rating update after every rated match.