Your DUPR rating comes with a second number that most players ignore, and it explains more of their frustration than the rating itself. It is the reliability score. If your rating will not move no matter how well you play, or if it swings wildly from one week to the next, the reliability score is usually the reason.
Here is what it means, how it is built, why it fades, and how to push it up.
What the reliability score is
Reliability is a separate figure from 1 to 100 percent that tells DUPR how confident it is that your rating reflects your true level. It is not a measure of how good you are. It is a measure of how much the system trusts the number. A 4.0 with 40 percent reliability and a 4.0 with 90 percent reliability have the same rating but very different standing: the algorithm treats the second one as settled and the first one as a work in progress.
Reliability level | What it means for your rating |
|---|---|
Below 60% | The rating moves slowly and good results barely show. Fix this before chasing the number. |
60% and above | The rating responds properly, and is raised by more matches, recent play, and varied opponents. |
The line to remember is roughly 60 percent. At or above it, DUPR considers your rating reliable. Below it, the algorithm moves your number cautiously, which is exactly why strong results sometimes fail to move the needle.
The three things that build reliability
Reliability rises with three inputs.
- Number of matches. More logged results give the algorithm more evidence. Most players cross into reliable territory somewhere around 10 to 20 matches.
- Recency. Recent matches count for more. A rating built on games from last year is less trusted than one built on games from last month.
- Variety of opponents. Playing a range of different, already-rated opponents builds reliability faster than playing the same two friends over and over.
If you want your rating to settle quickly, the move is obvious: play a good number of recent matches against varied, rated opponents. That is also the fastest path to a rating you can trust, which we cover in how to improve your DUPR.
Why reliability decays
Reliability is not permanent. Stop playing and it drifts down. A score sitting at 100 percent can fall back toward 60 percent over roughly six months of inactivity, because the recency input decays. The practical consequence catches a lot of players out: take a long break, come back for a tournament, and your rating is slow to respond because the system is no longer confident in it. In a smaller local scene it is easy to play hard for one event and then rest for months, which lets reliability go stale right before you need it.
Singles and doubles each have their own reliability
DUPR keeps separate reliability scores for singles and doubles, just as it keeps separate ratings. If you only ever play doubles, your singles reliability stays low and your singles rating stays unsettled. Neglect one format and its reliability quietly decays while the other stays healthy.
The connectivity catch
There is a subtler factor behind reliability: how well your local pool connects to the wider rated population. DUPR ratings are only meaningful when your cluster of players links back to established, rated opponents. If a group of unrated players only ever plays each other, their ratings float free of the global scale. The moment one of them plays a rated opponent, the whole group anchors to it. This is why a thin or isolated local scene produces lower reliability and slower stabilisation, a point that matters for newer markets and one we get into in getting rated in Malaysia.
How reliability affects your rating
Low reliability does not just sit there as a statistic. It actively dampens how much each result moves you. When the system is unsure of your level, it makes small, conservative adjustments so a couple of flukey games do not send your rating flying. That protects you from noise, but it also means a genuinely strong stretch of play takes longer to show up. If you are asking why your rating is stuck, the honest first answer is usually reliability, not the algorithm misjudging your games. The mechanics of how results are scored are in how the DUPR algorithm works.
Chasing a higher rating before your reliability is solid is like tuning an engine that is not warmed up. Get past 60 percent first with a run of recent, varied matches. Then the results you earn actually count.
Common questions
What is a good DUPR reliability score?
60 percent or higher is considered reliable, and 100 percent means full confidence. Below 60 percent your rating is still settling and moves cautiously.
Why is my DUPR reliability low?
Usually too few matches, matches that are not recent, or playing the same small group of opponents. Log more recent games against a variety of rated players to raise it.
Does my DUPR rating drop if I stop playing?
Your rating number does not automatically fall, but your reliability decays, drifting toward 60 percent over about six months of inactivity, which makes the rating slower to respond when you return.
How many matches until my rating is reliable?
Most players reach a reliable rating after 10 to 20 logged matches against varied opponents, though a thin local pool can slow that down. See what counts as a good rating for where you might land.
