ELO Rating System

1 Overview & Purpose

The Nexus ELO rating system provides a single, unified numerical measure of each player's competitive skill level across all rated competitions. It was designed to serve two primary purposes:

• Reflect current form — ratings evolve continuously with every match result, inactivity period, and division movement.
• Enable fair tournament seedings — a meaningfully comparable rating across the entire active player population.

The system is inspired by the original Elo system developed by Arpad Elo for chess and adapted widely across table tennis, including by the International Table Tennis Federation (ITTF).

2 Division Starting Ratings

Every player's rating is seeded from their division placement when they enter the system — not a generic default. This means ratings are calibrated to real competitive tiers from day one.

These starting points then evolve through actual match results. The gaps between divisions (200 points each at the top, tighter at the bottom) reflect the expected competitive difference between adjacent divisions.

3 K-Factor & Rating Changes

The K-factor controls the maximum points a player can gain or lose in a single match. The system uses a tiered K-factor based on current rating:

The same K-factor formula applies regardless of which league or format the game was played in — there is no "soft" local calculation that would create isolated rating pools.

4 Unified Rating Pool

All rated competitions feed into a single, shared rating table. There are no separate silos per league or format:

• CTTT League (all divisions)
• Summer League
• CTTT Junior League
• Super Vets League
• Open Tournaments (group stage and knockout)

This is one of the most important structural features of the system. A Premier player who beats a 5th Division player in an open tournament — that exchange is real, recorded, and affects both ratings in the same unified pool. Friendlies and practice matches are excluded from rating calculations.

5 Division Bumps on Promotion / Relegation

When a player is promoted or relegated, their rating is adjusted to better reflect their new competitive context. This prevents two systematic distortions:

• Without bumps on relegation: A relegated player would be too strong for their new division, systematically over-rewarding opponents who beat them (easy points against a high-rated player) and under-penalising opponents they beat.
• Without bumps on promotion: A promoted player would be systematically too weak — opponents face inflated losses from losing to someone whose rating doesn't yet reflect their true level.

Each division crossing creates a calibrated bridge point between the two competitive pools. Over multiple seasons of promotion and relegation across the player population, this creates a dense web of indirect rating connections that ties all divisions together.

6 Inactivity Decay

Ratings decay gradually when a player stops competing for an extended period. This serves two purposes: keeping individual ratings current, and preventing the overall rating pool from inflating due to stale high ratings anchoring ELO exchanges.

Grace Period & Rates

There is a 3-month grace period — no decay applies if a player has competed within the last 3 months. After that, decay is applied monthly:

Global Floor & Decay Cap

Two protections prevent runaway decay:

• Global floor (1100): No player's rating will drop below 1100 regardless of inactivity duration.
• 25% decay cap per inactive period: In any single inactive period (between returns to play), a player can lose at most 25% of their division's floor value. This resets when they return and compete again.

Every decay event is logged individually in the player's rating history with full transparency — visible in the ELO History tab on their player profile, with the applied weekly rate shown on hover.

7 Rating Pool Isolation

Rating pool isolation is a well-known concern in any ELO system. It occurs when separate groups of players rarely interact, creating rating "bubbles" that are internally consistent but not meaningfully comparable across groups. A player rated 1600 in one bubble could be genuinely stronger or weaker than a 1600 in another.

Structural Defences in This System

• Division-anchored starting ratings — ratings begin at values calibrated to real competitive tiers, establishing a consistent scale across the population from day one.
• Single unified pool — all leagues and tournaments feed the same rating table. There are no mechanical silos per competition type.
• Open tournaments as cross-pollination events — regular open tournaments mix players from all divisions and clubs in a single draw, creating direct cross-boundary ELO exchanges. These are the primary real-world mechanism for anchoring the pools together.
• Division bumps — every promotion and relegation creates a calibrated bridge point between adjacent division pools. Over multiple seasons, this connects all divisions through a dense network of adjusted ratings.
• Inactivity decay — prevents stale high ratings from inflating ELO exchanges for active players who happen to face returning long-term absentees.
• Universal K-factor — no format-specific K-factor that would allow separate rating economies to form in different competitions.

Residual Risks

• Division depth disparity: Direct cross-divisional games only occur at boundaries and in tournaments. Premier and 7th Division players rarely meet directly; the pools are linked through a chain, not direct contact. This is inherent to any tiered league structure.
• Satellite league isolation: Players who exclusively participate in Super Vets or Junior League and never enter a CTTT League match or open tournament could develop a slightly drifted sub-pool. This is a participation pattern issue, not a system design flaw.
• New player calibration period: Despite division-anchored seeds, a truly exceptional new player placed in the wrong division will take approximately 10–15 games to converge. The K=32 floor for lower-rated players accelerates this convergence.

8 Use in Tournament Seedings

The ELO system was designed with tournament seedings as a primary use case. Ratings are meaningfully comparable across the full active player population because:

• All competitive formats feed the same pool
• Starting ratings are calibrated to real competitive tiers
• Division bumps correct for tier transitions
• Inactivity decay keeps ratings current
• Open tournaments provide regular cross-divisional calibration

For the vast majority of tournament entrants who participate in league play and/or tournaments regularly, their ratings reflect genuine, comparable competitive ability.

Edge cases — brand new players, long-returning players with decayed ratings, or players who only participate in a single satellite league — are visible in the data and can be handled with Tournament Director discretion where necessary.

9 Overall Assessment

The Nexus ELO rating system is a robust, well-designed system for a regional club ecosystem. It addresses the most common weaknesses of club-level ELO implementations through several intentional design choices working together: