RTGDUPR-Style Rating

Formula

new = self + 0.05 × (actual − expected) + 0.02 × margin,  expected = 1 / (1 + 10^((opp − self) / 1.5))

Example

A 3.500-rated player beats a 4.000-rated opponent 11–8 in two games (margin +2): expected ≈ 0.32, actual = 1. New rating ≈ 3.500 + 0.05 × 0.68 + 0.04 = 3.574.

Why it matters

Win percentage rewards farming weak opponents. Rating rewards genuine difficulty — a 5-and-5 record against tougher players is worth more than 10-and-0 against the bottom of the bracket.

In depth

ArcStat ships a DUPR-style rating: a single number on the 2.0–8.0 scale that represents a player’s skill independent of which league they happen to be in. Every match nudges the rating up or down based on the opponent’s rating and the margin of the result.

It is not real DUPR — DUPR is a registered service of Dynamic Universal Pickleball Rating. ArcStat’s implementation is the same shape (Elo-style update, margin-weighted) so the numbers feel familiar to players who already know their DUPR.

Singles and doubles ratings are tracked separately. A 4.500 singles player is often a 4.000 doubles player and vice versa.