How to Rank Ping Pong Players Fairly
6 min read
Ping pong is the perfect game for a rating system: matches are short, you play a lot of them, and skill gaps show up fast. But a simple win-loss record falls apart the moment people play different opponents. Here’s how to rank table tennis players in a way everyone will accept as fair.
Why win-loss records don’t work
Imagine Maya goes 8–2, all against beginners, while Sam goes 6–4 against the office sharks. Maya’s record looks better, but Sam is clearly the stronger player. Raw win percentage rewards whoever picks the easiest opponents. To rank fairly, you need a system that accounts for who you beat, not just how often. That’s exactly what ELO does.
Setting up a ping pong ELO ladder
- Start every player at the same rating (1000 is a clean default).
- Use a K-factor of 32 — ping pong is high-volume, so ratings settle quickly and 32 keeps things lively. See how to choose a K-factor if you want to fine-tune.
- Record the winner and loser of every game. That’s the whole job.
- Let the ratings sort everyone out over the first couple of weeks.
Best-of-1, best-of-3, or best-of-5?
Single games are noisy — the weaker player wins a one-off more often than you’d think. For ranking purposes, decide on one format and stick to it:
- Single games to 11: fast, lots of data, more variance per match. Fine when you play often.
- Best of 3 or 5: more reliable per result, fewer fluke outcomes, but slower to accumulate matches.
Record one ELO result per match (the overall winner), not per individual game within a best-of series, so the format stays consistent.
Should the score margin count?
By default an 11–9 squeaker and an 11–2 thrashing move the same points. If your group feels dominance should matter, turn on score-margin scoring and set the target score to 11. Blowouts will then move more points than close games. Most casual groups leave this off — it’s one less thing to argue about — but it’s a nice touch for competitive clubs.
Handling doubles
If you play 2-on-2, treat each pairing as a “team” and rate the teams, or keep doubles in a separate league from singles. Mixing singles and doubles results into one rating muddies the signal, because doubles skill and singles skill aren’t the same thing.
Keeping it fair over time
- Encourage varied matchups. Ratings are most accurate when everyone plays a spread of opponents, not just their two desk neighbors. A random-matchup feature helps here.
- Don’t reset too often. Wiping ratings every month throws away hard-won information. An annual reset, if any, is plenty.
- Cap nothing artificially. Let the strong players be highly rated; that’s the point. The gap is the information.
Run it for a few weeks and the ladder will reflect real pecking order far better than any “I beat you that one time” memory. Cap it off with a single- or double-elimination tournament when you want a champion — see how to seed a bracket fairly.